# Iguazio

> Keep up to date on the latest trends in AI, genAI, data science & MLOps (operationalizing machine learning).

---


## Pages

- [De-Risk Your Gen AI Applications](https://www.iguazio.com/solutions/de-risk-your-gen-ai-applications/): Eliminate LLM risks for Gen AI apps. Meet compliance, reduce toxicity and biad and ensure peak model performance.
- [Company](https://www.iguazio.com/company/): We’re building a faster way for data science to flow seamlessly from raw idea to real-world impact. This is data science brought to life.
- [Gen AI Ops](https://www.iguazio.com/mlops-for-generative-ai/): Generative AI Ops streamlines the way you operationalize and scale your generative AI models with an approach that focuses on business value.
- [CI/CD for ML](https://www.iguazio.com/solutions/ci-cd-for-ml/): Automate and simplify the building, testing, deployment and monitoring of AI=, Continuously train, test and deploy your LLMs and ML models
- [AI in Secure IT Environments: AWS GovCloud & SCIF](https://www.iguazio.com/secure-it-environments-aws-govcloud/): The Iguazio AI Platform allows data science teams to run their machine learning pipelines on AWS GovCloud or in a SCIF.
- [Data Mesh](https://www.iguazio.com/solutions/data-mesh-for-mlops/): Implement a Data Mesh approach to data architecture to make data accessible, interconnected and valuable across the organization
- [Gaming](https://www.iguazio.com/solutions/gaming/): Use data science to optimize gaming experiences. Anticipate users' next move, reduce church, increase ROI and provide the best user experience possible.
- [Security](https://www.iguazio.com/security/): Securely develop, deploy and manage real-time AI applications at scale. Secure Environment for Development.
- [ODSC MLOps Resource Center](https://www.iguazio.com/mlops-resource-center/): Thanks for meeting with us! Enjoy the full MLOps resource center below. Find here the best resources from our experts to you.
- [Technology](https://www.iguazio.com/technology/): The Iguazio AI Platform enables you to develop, deploy and manage real-time AI applications at scale.
- [Iguazio’s ESG Strategy](https://www.iguazio.com/esg-strategy/): Iguazio is fully committed to sustainability and to helping our customers and partners make real-world impact where it matters most.
- [Questions](https://www.iguazio.com/questions/): Find here questions about data science, machine learning and serverless automation and meet Iguazio experts at the industry's leading events
- [Energy and Utilities](https://www.iguazio.com/solutions/energy-and-utilities/): Energy and utilities operate in data-rich environments of customer and sensor data which can be leveraged using AI to transform the way they do business.
- [Hackathon Terms](https://www.iguazio.com/hackathon-terms/): Find here the MLOps for Good Hackathon Terms and Conditions, read everything and contact us in case of doubts and suggestions.
- [MLOps Glossary](https://www.iguazio.com/glossary/): This glossary will help you understand the main terms, acronyms, and abbreviations related to AI apps at scale in an end-to-end MLOps platform.
- [Technology OLD](https://www.iguazio.com/technology-old-2/): The Iguazio Data Science Platform enables you to develop, deploy and manage real-time AI applications at scale.
- [MLRun](https://www.iguazio.com/open-source/mlrun/): Open source AI orchestration open source framework, to run code either locally on your PC for or on a large scale Kubernetes cluster.
- [MLOps](https://www.iguazio.com/mlops/): MLOps is the practice of creating continuous development, integration and delivery (CI/CD) of data and ML intensive applications.
- [Real-Time Feature Engineering](https://www.iguazio.com/real-time-feature-engineering/): Feature engineering made simple: Ingest real-time data, perform data transformation, generate and share real-time features across teams.
- [Model Monitoring](https://www.iguazio.com/solutions/model-monitoring/): Continuously track models in production to automatically detect drift and maintain accuracy in rapidly changing live environments
- [Healthcare](https://www.iguazio.com/solutions/healthcare/): The Iguazio Data Science Platform for Healthcare enables healthcare facilities to develop, deploy and manage AI applications.
- [Home - New](https://www.iguazio.com/): The Iguazio AI platform operationalizes and de-risks ML & gen AI applications at scale. Turn your AI projects into real business impact.
- [Customers](https://www.iguazio.com/customers/): Transforming AI initiatives into real-world outcomes with MLOps. Learn how customers use the Iguazio MLOps Platform.
- [Integrated Feature Store](https://www.iguazio.com/feature-store/): Engineer online and offline features, and make them accessible to everyone with one fully integrated feature store.
- [What Is A Machine Learning Pipeline?](https://www.iguazio.com/machine-learning-pipeline/): A machine learning pipeline helps to streamline and speed up the process by automating these workflows and linking them together.
- [What Is Enterprise Data Science?](https://www.iguazio.com/enterprise-data-science/): Enterprise data science combines data scientists, data engineers, IT teams, and more to generate value out of big data. 
- [What Are Machine Learning Pipeline Tools?](https://www.iguazio.com/machine-learning-pipeline-tools/): A machine learning pipeline tool helps automate and streamline machine learning pipelines. Learn more on our page.
- [Machine Learning Operations (MLOps)](https://www.iguazio.com/machine-learning-operations-mlops/): Machine learning operations (MLOps) is considered to be the backend supporting ML applications in business.
- [What is Operationalizing Machine Learning?](https://www.iguazio.com/operationalizing-machine-learning/): Operationalizing machine learning is one of the final stages before deploying and running an ML model in a production environment.
- [MLOps Live Webinar Series](https://www.iguazio.com/mlops-live-webinar-series/): Webinars with industry leaders sharing practical advice & demonstrating how they’ve made real business impact by bringing data science to life.
- [Iguazio Support Policy](https://www.iguazio.com/supportpolicy/): Explore Iguazio's support policy for seamless assistance. Find out how we are dedicated to helping you every step of the way.
- [Customer Support](https://www.iguazio.com/support/): Iguazio’s dedicated specialists are here to support you in deploying real-world data science applications for business impact.
- [Terms of Use](https://www.iguazio.com/terms-of-use/): Discover the comprehensive terms of use for the Iguazio platform. Ensure safe & compliant use by acquainting yourself with our policies.
- [Privacy Policy](https://www.iguazio.com/privacy-policy/): Understand your privacy rights on the Iguazio wesbite. Learn how we protect your data and prioritize your privacy.
- [Career Inner Page](https://www.iguazio.com/career-inner-page/): We’re building the future of data science. Join our team of experts, see our open positions.
- [AI Pipeline Orchestration](https://www.iguazio.com/solutions/pipeline-orchestration/): Manage your gen AI and ML workflow at scale to get from pilot to production faster. Hybrid and on prem deployment. Automations.
- [News & Events](https://www.iguazio.com/news-events/): Read the latest news on Iguazio and meet our experts at the industry's leading events for data science, ML and AI.
- [Nuclio](https://www.iguazio.com/open-source/nuclio/): Discover Nuclio for seamless serverless pipeline automation, enhancing scalability and performance for your data applications.
- [Careers](https://www.iguazio.com/careers/): We're building the future of AI. If you think you're the perfect fit for Iguazio send us your CV. Join Us!
- [Serverless Automation](https://www.iguazio.com/solutions/serverless-automation/): Automate each step of the pipeline with enterprise performance, scale and reliability using Nuclio, the open source serverless framework
- [Telecommunications](https://www.iguazio.com/solutions/telecommunications/): Create automatically self-healing systems, improving network optimization, churn reduction, customer engagement and more.
- [Manufacturing](https://www.iguazio.com/solutions/manufacturing/): Enhance manufacturing efficiency with Iguazio's AI platform. Real-time insights and AI-driven solutions.
- [Retail](https://www.iguazio.com/solutions/retail/): Transform retail with Iguazio's AI-driven platform. Boost sales, optimize operations, and enhance customer experience.
- [Smart Mobility](https://www.iguazio.com/solutions/smart-mobility/): The Iguazio AI platform for Smart Mobility enables you to operationalize and de-risk AI applications.
- [Ad-Tech](https://www.iguazio.com/solutions/ad-tech/): Use data science to optimize ad-tech. Anticipate users' next move, increase engagement and provide the best user experience possible.
- [Solutions](https://www.iguazio.com/solutions/): Leverage artificial intelligence to power true business impact with AI across industries, technologies and use-cases.
- [GPU Management](https://www.iguazio.com/solutions/gpu-as-a-service/): Accelerate AI, deep learning and data processing using a GPU provisioning solution, for faster and scalable AI-based applications
- [Open Source](https://www.iguazio.com/open-source/): Iguazio's open source projects for Real-time Serverless Functions and ML Pipeline Orchestration
- [Partners](https://www.iguazio.com/partners/): Join our channel and technology partner programs to work with a pioneering data science platform and benefit from Iguazio's growing partner ecosystem.
- [Contact](https://www.iguazio.com/contact/): Have a question or just need more info about Iguazio? Contact us and we’ll get back to you!
- [Financial Services](https://www.iguazio.com/solutions/financial-services/): From mitigating risks proactively, to creating personalized customer services, harness the power of AI to transform your bottom line.
- [AI Blog](https://www.iguazio.com/blog/): Keep up to date on the latest trends in AI, genAI, data science & MLOps (operationalizing machine learning).
- [Platform](https://www.iguazio.com/platform/): Meet the platform that automates AI and cuts the time to impact of your gen AI and ML applications. Click for more.

---


## Posts

- [13 Best Free Datasets for Call Centers and Telcos](https://www.iguazio.com/blog/13-best-free-datasets-for-call-centers-and-telcos/): Here are 13 open datasets and data sources for telcos and call centers, that you can use for (gen) AI projects
- [Preventing Data Leakage in Gen AI Chatbots: What’s Your Risk Appetite?](https://www.iguazio.com/blog/preventing-data-leakage-in-gen-ai-chatbots-whats-your-risk-appetite/): Chatbots are not just a tech choice, they're also a new security risk. The right guardrails depends on your organizational risk appetite.
- [Build Observable Data Flywheels for Production with Iguazio’s MLRun and NVIDIA NeMo Microservices](https://www.iguazio.com/blog/observable-data-flywheels-for-production-with-mlrun-and-nvidia-nemo/): We are proud to announce a new integration between MLRun and NVIDIA NeMo microservices, by extending NVIDIA Data Flywheel Blueprint.
- [Deploying Gen AI in Production with NVIDIA NIM & MLRun](https://www.iguazio.com/blog/deploying-gen-ai-in-production-with-nvidia-nim-mlrun/): How to use NVIDIA NIM with MLRun to build production-ready, scalable and high-performing gen AI applications.
- [MLRun v1.8 Now Available: Smarter Model Monitoring, Alerts and Tracking](https://www.iguazio.com/blog/mlrun-v1-8-release/): MLRun v1.8 adds features to make LLM and ML evaluation and monitoring more accessible, practical and resource-efficient.
- [The Future of AI Monitoring: How to Address a Non-Negotiable, Yet Still Developing, Requirement ](https://www.iguazio.com/blog/the-future-of-ai-monitoring-how-to-address-a-non-negotiable-yet-still-developing-requirement/): Here are our basic principles for designing a monitoring system to ensure gen AI applications are predictable, reliable, and safe at scale.
- [Building Agent Co-pilots for Proactive Call Centers](https://www.iguazio.com/blog/building-agent-co-pilots-for-proactive-call-centers/): How a large health insurance provider implemented an agentic co-pilot designed scale across multiple call centers and environments.
- [Best 10 Free Datasets for Manufacturing [UPDATED]](https://www.iguazio.com/blog/free-manufacturing-datasets/): Here are 10 excellent open manufacturing datasets and data sources for manufacturing data for machine learning. Get now.
- [11 Best Free Retail Datasets for Machine Learning [UPDATED]](https://www.iguazio.com/blog/13-best-free-retail-datasets-for-machine-learning/): Retail datasets for ML can be hard to find. Here are 17 excellent open retail datasets and data sources for your next ML project.
- [How to Manage Thousands of Real-Time Models in Production](https://www.iguazio.com/blog/how-to-manage-thousands-of-real-time-models-in-production/): Learn how to efficiently manage thousands of real-time models in production with strategies to ensure scalability, accuracy, and performance.
- [Introducing Agentic RAG: The Best of Both Worlds](https://www.iguazio.com/blog/introducing-agentic-rag-the-best-of-both-worlds/): How Agentic RAG revolutionizes AI-powered applications by making them more autonomous, intelligent, and context-aware.
- [Gen AI Trends and Scaling Strategies for 2025](https://www.iguazio.com/blog/gen-ai-trends-and-scaling-strategies-for-2025/): How should enterprises respond to the turbo charged changes in gen AI trends? Gartner's strategies for the year ahead.
- [AI Agent Training: Essential Steps for Business Success](https://www.iguazio.com/blog/ai-agent-training-essential-steps-for-business-success/): How to train AI agents that are not only powerful but also reliable, scalable, and aligned with business goals. Read more here.
- [Best 13 Free Financial Datasets for Machine Learning [Updated]](https://www.iguazio.com/blog/best-13-free-financial-datasets-for-machine-learning/): Developing ML models for financial or economic use cases? Here are 13 great open financial datasets to develop and train ML models.
- [Gen AI or Traditional AI: When to Choose Each One](https://www.iguazio.com/blog/gen-ai-or-traditional-ai-when-to-choose-each-one/): How do you decide when to use traditional AI vs. generative AI? Here's our roadmap to achieving your business goals with each.
- [Top Gen AI Demos of AI Applications With MLRun](https://www.iguazio.com/blog/top-gen-ai-demos-of-ai-applications-with-mlrun/): A roundup of our top gen AI demo videos showing how to build and manage AI applications with open-source MLRun.
- [6 Best Practices for Implementing Generative AI](https://www.iguazio.com/blog/6-best-practices-for-implementing-generative-ai/): This guide outlines six best practices to ensure your generative AI initiatives are effective: valuable, scalable, compliant and future-proof.
- [2025 Gen AI Predictions: What Lies Ahead?](https://www.iguazio.com/blog/2025-gen-ai-predictions-what-lies-ahead/): Cheers to a successful 2025! Who will thrive in this new ecosystem? Here are my predictions for the upcoming year.
- [Choosing the Right-Sized LLM for Quality and Flexibility: Optimizing Your AI Toolkit](https://www.iguazio.com/blog/choosing-the-right-sized-llm-for-quality-and-flexibility-optimizing-your-ai-toolkit/): Are you using an LLM that is fit to your specific use case? Here's our guide on choosing and optimizing LLMs for performance and cost.
- [MLRun v1.7 Launched — Solidifying Generative AI Implementation and LLM Monitoring](https://www.iguazio.com/blog/mlrun-v1-7-launched-solidifying-generative-ai-implementation-and-llm-monitoring/): MLRun 1.7 is now available with powerful features for GenAI implementation, with a special emphasis on LLM monitoring.
- [Gen AI for Marketing - From Hype to Implementation](https://www.iguazio.com/blog/gen-ai-marketing-hype-implementation/): Gen AI for marketing use cases - how? Read to learn a staged approach for rolling out gen AI, use cases, a demo & examples
- [Implementing Gen AI in Regulated Sectors: Finance, Telecom, and More](https://www.iguazio.com/blog/how-to-implement-gen-ai-in-highly-regulated-environments/): Gen AI challenges are exacerbated in highly-regulated industries, such as financial services and telecommunications. Here's what to do.
- [Building Scalable Gen AI Apps with Iguazio and MongoDB](https://www.iguazio.com/blog/building-gen-ai-applications-with-iguazio-and-mongodb/): Discover how to build & scale gen AI apps with Iguazio (acquired by McKinsey) and MongoDB - with simplicity, performance & risk mitigation
- [RAG vs Fine-Tuning: Navigating the Path to Enhanced LLMs](https://www.iguazio.com/blog/rag-vs-fine-tuning/): RAG uses external resources to train models, fine-tuning involves further training on specialized datasets. Learn when to use each one and why
- [Commercial vs. Self-Hosted LLMs: A Cost Analysis & How to Choose the Right Ones for You](https://www.iguazio.com/blog/commercial-vs-self-hosted-llms/): How to choose a proprietary LLM vs. self-hosting open-source. See use cases, requirements and benchmarks, and optimization tools for LLMs.
- [Transforming Enterprise Operations with Gen AI](https://www.iguazio.com/blog/transforming-enterprise-operations-with-gen-ai/): Effectively implement and scale gen AI while avoiding risk. See use cases, from R&D to automotive to the supply chain.
- [Future-Proofing Your App: Strategies for Building Long-Lasting Apps](https://www.iguazio.com/blog/future-proofing-your-gen-ai-aplications-strategies-for-building-long-lasting-apps/): Learn to build gen AI pipelines that are modular, so they can support up-to-date LLM deployment and management
- [LLM Validation and Evaluation](https://www.iguazio.com/blog/llm-validation-and-evaluation/): LLM validation & evaluation is about assessing the performance and capabilities. Learn LLM methods and see a demo with crowd sourcing.
- [Integrating LLMs with Traditional ML: How, Why & Use Cases](https://www.iguazio.com/blog/integrating-llms-with-traditional-ml-how-why-use-cases/): Integrating LLMs with traditional ML models enhances each model’s capabilities. Discover the benefits of integration and example use cases.
- [LLM Metrics: Key Metrics Explained](https://www.iguazio.com/blog/llm-metrics-key-metrics-explained/): The top LLM metrics to measure for higher performing models at higher efficiency, while ensuring privacy and eliminating bias and toxicity.
- [Generative AI in Call Centers: How to Transform and Scale Superior Customer Experience](https://www.iguazio.com/blog/generative-ai-in-call-centers-how-to-transform-and-scale-superior-customer-experience/): The coming transformative impact of gen AI on call centers, and how to build a gen AI call center analysis app now.
- [Why You Need GPU Provisioning for GenAI](https://www.iguazio.com/blog/why-you-need-gpu-as-a-service-for-genai/): GPU provisioning simplifies the management of GPUs to improve performance and save significant costs. Here's how and why to leverage
- [Implementing Gen AI for Financial Services](https://www.iguazio.com/blog/implementing-gen-ai-for-financial-services/): How Gen AI revolutionizes financial services, its potential impact, pitfalls, and strategies. Insights from industry experts included.
- [LLMOps vs. MLOps: Understanding the Differences](https://www.iguazio.com/blog/llmops-vs-mlops-understanding-the-differences/): LLMOps applies MLOps principles to LLMs. This post delves into the concepts of LLMOps and MLOps, explaining how and when to use each one.
- [Implementing Gen AI in Practice](https://www.iguazio.com/blog/implementing-genai-in-practice/): How to build an ‘AI Factory’ that streamlines and simplifies the process of rolling out new generative AI applications. Read more here.
- [How HR Tech Company Sense Scaled their ML Operations using Iguazio](https://www.iguazio.com/blog/how-hr-tech-company-sense-scaled-their-ml-operations-using-iguazio/): Customer story: How Sense built their AI chatbot using a complex NLP serving pipeline, and how they overcame some ML challenges with Iguazio.
- [What Lays Ahead in 2024? AI/ML Predictions for the New Year](https://www.iguazio.com/blog/ai-ml-predictions-for-the-new-year/): Iguazio CTO Yaron Haviv's predictions for the upcoming year: What will be different in the way businesses approach generative AI this year?
- [16 Best Free Human Annotated Datasets for Machine Learning](https://www.iguazio.com/blog/best-free-human-annotated-datasets-for-ml/): Human-annotated datasets offer a level of precision, nuance, and contextual understanding that automated methods struggle to match.
- [Introducing our New Book: Implementing MLOps in the Enterprise](https://www.iguazio.com/blog/introducing-our-new-book-implementing-mlops-in-the-enterprise/): A practical guide which guides IT leaders to bring data science to life for a variety of real-world MLOps scenarios. Find here.
- [Scaling MLOps Infrastructure: Components and Considerations for Growth](https://www.iguazio.com/blog/scaling-mlops-infrastructure-components-and-considerations-for-growth/): When it comes to scaling your MLOps operations, a high-quality, reliable and effective MLOps platform is essential for growth.
- [How to Build a Smart GenAI Call Center App](https://www.iguazio.com/blog/how-to-build-a-smart-genai-call-center-app/): Building a generative AI smart call center app is a promising solution for improving the customer experience and call center efficiency.
- [Top 27 Free Healthcare Datasets for Machine Learning](https://www.iguazio.com/blog/top-22-free-healthcare-datasets-for-machine-learning/): We pulled together 27 excellent open datasets in the field of healthcare for your next machine learning project. Learn more about it here.
- [Top 10 ODSC West Sessions You Must Attend in 2023](https://www.iguazio.com/blog/odsc-west-sessions-you-must-attend/): Planning to attend ODSC West 2023? We've compiled the top sessions at this conference that we're most looking forward to.
- [28 Best Free NLP Datasets for Machine Learning](https://www.iguazio.com/blog/nlp-datasets/): Building an AI application with NLP? You'll need a robust dataset. Here are some of the top open NLP datasets for you to leverage.
- [How to Mask PII Before LLM Training](https://www.iguazio.com/blog/how-to-mask-pii-before-llm-training/): The PII Recognizer is an open source function that can detect and anonymize PII data in datasets, so you can build Gen AI apps securely.
- [Model Observability and ML Monitoring: Key Differences and Best Practices](https://www.iguazio.com/blog/model-observability/): We delve into the distinctions between model observability and ML monitoring, shedding light on their unique attributes and functionalities.
- [Implementing MLOps: 5 Key Steps for Successfully Managing ML Projects](https://www.iguazio.com/blog/implementing-mlops-5-key-steps-for-successfully-managing-ml-projects/): MLOps accelerates the ML deployment process. Here are the critical steps of MLOps and what to look for in an MLOps platform.
- [MLOps for Generative AI in the Enterprise](https://www.iguazio.com/blog/mlops-for-generative-ai-in-the-enterprise/): How to leverage LLMs in live enterprise applications, and embed Responsible AI principles into the process. Find more here.
- [Mastering ML Model Performance: Best Practices for Optimal Results](https://www.iguazio.com/blog/mastering-ml-model-performance-best-practices-for-optimal-results/): Evaluating ML model performance is essential for ensuring the reliability, quality, accuracy and effectiveness of your ML models.
- [What are the Advantages of Automated Machine Learning Tools?](https://www.iguazio.com/blog/automl-advantages/): The benefits of AutoML, the available tools and how to choose the right tool for your industry, plus how to implement AutoMLOps
- [Integrating MLOps with MLRun and Databricks](https://www.iguazio.com/blog/integrating-mlops-with-mlrun-and-databricks/): How to leverage multiple MLOps tools to streamline model serving for complex real-time use cases. Find here.
- [Deploying Machine Learning Models for Real-Time Predictions Checklist](https://www.iguazio.com/blog/deploying-machine-learning-models-for-real-time-predictions-checklist/): Here are our recommendations for data professionals who want to improve and streamline their real-time model deployment process.
- [Top 7 ODSC East Sessions You Can’t Afford to Miss](https://www.iguazio.com/blog/odsc-east-2023/): We've compiled the top sessions at ODSC East Boston 2023 that we're most looking forward to, covering topics like real-time ML, LLMs, data privacy and more.
- [Kubeflow Vs. MLflow Vs. MLRun: Which One is Right for You?](https://www.iguazio.com/blog/kubeflow-vs-mlflow-vs-mlrun/): We dive into these three tools to better understand their capabilities, and how they fit into the ML lifecycle. Learn more here.
- [How Seagate Runs Advanced Manufacturing at Scale With Iguazio](https://www.iguazio.com/blog/how-seagate-runs-advanced-manufacturing-at-scale-with-iguazio/): How Seagate tackled their predictive manufacturing use case with continuous data engineering at scale, keeping costs low and productivity high.
- [McKinsey Acquires Iguazio: Our Startup’s Journey](https://www.iguazio.com/blog/mckinsey-acquires-iguazio-our-startups-journey/): When I founded Iguazio with my co-founders, I never thought I would be making this announcement on our company blog: McKinsey acquired Iguazio!
- [HCI’s Journey to MLOps Efficiency](https://www.iguazio.com/blog/hcis-journey-to-mlops-efficiency/): Deploying ML in the enterprise is a complex process. Jiri Steuer from HCI shares his top tips and ideas for achieving MLOps efficiency.
- [Distributed Feature Store Ingestion with Iguazio, Snowflake, and Spark ](https://www.iguazio.com/blog/distributed-feature-store-ingestion-with-iguazio-snowflake-and-spark/): Our step-by-step guide on how to leverage distributed ingestion into the Iguazio feature store with Snowflake and Spark for ingestion.
- [Looking into 2023: Predictions for a New Year in MLOps](https://www.iguazio.com/blog/mlops-predictions-for-2023/): As we raise our glasses to the upcoming year, here are my predictions of what we'll see in the MLOps industry in 2023
- [Iguazio Named a Major Player in the IDC MLOps MarketScape 2022](https://www.iguazio.com/blog/idc-mlopmarketscape-2022/): IDC MarketScape is based on a rigorous framework that highlights the factors expected to be the most influential in the both short- and long term.
- [Iguazio Named a Leader and Outperformer In GigaOm Radar for MLOps 2022](https://www.iguazio.com/blog/iguazio-named-a-leader-and-outperformer-in-gigaom-radar-for-mlops-2022/): Iguazio is thrilled to be named an Outperforming Leader in GigaOm’s latest 2022 report on MLOps.
- [Deploying Your Hugging Face Models to Production at Scale with MLRun](https://www.iguazio.com/blog/deploying-your-hugging-face-models-to-production-at-scale-with-mlrun/): Here's how to continuously deploy Hugging Face models into real business environments at scale, along with the required application logic.
- [Top 10 ODSC West Sessions You Must Attend!](https://www.iguazio.com/blog/top-10-odsc-west-sessions-you-must-attend/): We've compiled the sessions at ODSC West 2022 in San Francisco that we're most looking forward to, covering a wide range of topics.
- [How to Run Workloads on Spark Operator with Dynamic Allocation Using MLRun](https://www.iguazio.com/blog/how-to-run-workloads-on-spark-operator-with-dynamic-allocation-using-mlrun/): With Spark Operator, dynamic executor allocation saves big costs. Here's how to abstract away the implementation complexities with MLRun.
- [Building an Automated ML Pipeline with a Feature Store Using Iguazio & Snowflake](https://www.iguazio.com/blog/building-an-automated-ml-pipeline-with-a-feature-store-using-iguazio-snowflake/): Here's how to activate your Snowflake data by using the Iguazio feature store to build, store and share features from your Snowflake data.
- [Iguazio Product Update: Optimize Your ML Workload Costs with AWS EC2 Spot Instances](https://www.iguazio.com/blog/iguazio-product-update-optimize-your-ml-workload-costs-with-aws-ec2-spot-instances/): Choosing a spot instance is a cost-saving choice if you can be flexible about when your applications run and if can be interrupted.
- [From AutoML to AutoMLOps: Automated Logging & Tracking of ML](https://www.iguazio.com/blog/from-automl-to-automlops/): Here's how to embrace AutoMLOps, to automate ML engineering tasks so that your code is automatically ready for production.
- [How to Deploy an MLRun Project in a CI/CD Process with Jenkins Pipeline ](https://www.iguazio.com/blog/how-to-deploy-an-mlrun-project-in-a-ci-cd-process-with-jenkins-pipeline/): In this article, we will walk you through steps to run a Jenkins server in docker and deploy the MLRun project using Jenkins pipeline. 
- [Beyond Hyped: Iguazio Named in 8 Gartner Hype Cycles for 2022](https://www.iguazio.com/blog/beyond-hyped-iguazio-named-in-8-gartner-hype-cycles-for-2022/): We’re very excited to share that Iguazio has been named a sample vendor in eight Gartner Hype Cycles for 2022.
- [Build an AI App in Under 20 Minutes](https://www.iguazio.com/blog/build-an-ai-app-in-under-20-minutes/): Here's how to build simple AI applications that leverage pre-built ML models and allow you to interact with a UI to visualize the results.
- [Machine Learning Experiment Tracking from Zero to Hero in 2 Lines of Code ](https://www.iguazio.com/blog/machine-learning-experiment-tracking-from-zero-to-hero-in-2-lines-of-code/): Here's how to turn your existing model training code into an MLRun job and get the benefit of all the experiment tracking, plus more. 
- [Iguazio Recognized in Gartner's 2022 Market Guide for DSML Engineering Platforms](https://www.iguazio.com/blog/gartner-2022-market-guide-for-dsml-engineering-platforms/): We’re proud to share that Iguazio has been included in Gartner's 2022 Market Guide for DSML Engineering Platforms.
- [The Easiest Way to Track Data Science Experiments with MLRun](https://www.iguazio.com/blog/experiment-tracking/): Learn how to solve experiment tracking complexity concerns with MLRun, a new open source framework which optimizes the management of ML operations.
- [Top 9 ODSC Europe Sessions You Can’t Miss!](https://www.iguazio.com/blog/top-9-odsc-europe-sessions-you-cant-miss/): See you soon at ODSC Europe in London! Here's our list of the top recommended sessions we're looking forward to attending.
- [Best Practices for Succeeding with MLOps ](https://www.iguazio.com/blog/best-practices-for-succeeding-with-mlops/): In our latest webinar, Noah Gift sat down with us to discuss best practices for succeeding with MLOps, and much more.
- [Top 8 Recommended MLOps World 2022 Sessions](https://www.iguazio.com/blog/top-8-recommended-mlops-world-2022-sessions/): We've compiled the top sessions at MLOps World 2022 in Toronto that we're most looking forward to. See you there!
- [Using Snowflake and Dask for Large-Scale ML Workloads](https://www.iguazio.com/blog/using-snowflake-and-dask-for-large-scale-ml-workloads/): Snowflake’s Connector for Python is a great fit for large-scale ML workloads with Dask. Here’s how to use it.
- [ODSC East Boston 2022 - Top 11 Sessions for AI and ML Professionals to Attend](https://www.iguazio.com/blog/odsc-east-boston-2022-top-11-sessions-for-ai-and-ml-professionals-to-attend/): ODSC East Boston is fast approaching! Our picks for the top sessions covering AI/ML, MLOps, ML use cases, AI explainability, and more.
- [Real-Time Streaming for Data Science](https://www.iguazio.com/blog/real-time-streaming-for-data-science/): Explore real-time streaming for data science with serverless architecture and time series analysis to drive smarter insights.
- [GigaOm Names Iguazio a Leader and Outperformer for 2022](https://www.iguazio.com/blog/gigaom-names-iguazio-a-leader-and-outperformer-for-2022/): GigaOm gave Iguazio top scores on several evaluation metrics in the GigaOm Radar for Data Science Platforms 2022 report.
- [Iguazio named in Forrester's Now Tech: AI/ML Platforms](https://www.iguazio.com/blog/iguazio-named-in-forresters-now-tech-ai-ml-platforms-q1-2022/): We are delighted to share that Iguazio has been named in Forrester’s Overview of the leading AI/ML Platform Providers
- [Top 8 Machine Learning Resources for Data Scientists, Data Engineers and Everyone](https://www.iguazio.com/blog/top-8-machine-learning-resources-for-data-scientists-data-engineers-and-everyone/): Explore our list of the top ML and MLOps learning resources, with blogs, video series, online communities and more.
- [ML Workflows: What Can You Automate?](https://www.iguazio.com/blog/ml-workflows-what-can-you-automate/): How automating various steps in the ML workflow using a MLOps approach can help data teams achieve faster deployment of ML models
- [Orchestrating ML Pipelines at Scale with Kubeflow](https://www.iguazio.com/blog/orchestrating-ml-pipelines-scale-kubeflow/): It’s okay if you’re a hobbyist, but data science models are meant to be incorporated into real business applications. Read more here.
- [Automating MLOps for Deep Learning: How to Operationalize DL With Minimal Effort](https://www.iguazio.com/blog/automating-mlops-for-deep-learning-how-to-operationalize-dl-with-minimal-effort/): How to use MLRun to orchestrate and automate the process of taking deep learning pipelines from research to production.
- [What Are Feature Stores and Why Are They Critical for Scaling Data Science?](https://www.iguazio.com/blog/what-are-feature-stores-and-why-are-they-critical-for-scaling-data-science/): Feature stores enable data scientists to reuse features instead of rebuilding these features again for different models, saving them valuable time.
- [The Complete Guide to Using the Iguazio Feature Store with Azure ML - Part 4](https://www.iguazio.com/blog/the-complete-guide-to-using-the-iguazio-feature-store-with-azure-ml-part-4/): Part 4 of our 4-part blog series that will take you through how the Iguazio feature store works with Microsoft Azure Cloud.
- [The Complete Guide to Using the Iguazio Feature Store with Azure ML - Part 3](https://www.iguazio.com/blog/the-complete-guide-to-using-the-iguazio-feature-store-with-azure-ml-part-3/): Part 3 of a 4-part blog series that will take you through how the Iguazio feature store works with Microsoft Azure Cloud.
- [The Complete Guide to Using the Iguazio Feature Store with Azure ML - Part 2](https://www.iguazio.com/blog/the-complete-guide-to-using-the-iguazio-feature-store-with-azure-ml-part-2/): Part 2 of a 4-part blog series that will take you through how the Iguazio feature store works with Microsoft Azure Cloud.
- [The Complete Guide to Using the Iguazio Feature Store with Azure ML - Part 1](https://www.iguazio.com/blog/part-one-the-complete-guide-to-using-the-iguazio-feature-store-with-azure-ml/): This 4-part blog series will take you through how the Iguazio feature store works with Microsoft Azure Cloud.
- [Looking into 2022: Predictions for a New Year in MLOps](https://www.iguazio.com/blog/2022-predictions/): Before we toast to a new year ahead, here are my predictions of what awaits the MLOps industry in 2022.
- [Adopting a Production-First Approach to Enterprise AI](https://www.iguazio.com/blog/adopting-a-production-first-approach-to-enterprise-ai/): Modern AI applications require a continuous operational pipeline and a production-first approach to make it all feasible.
- [Introduction to TF Serving](https://www.iguazio.com/blog/introduction-to-tf-serving/): Our in-depth guide to what Tensorflow Serving is, why you need it, and how to use it, for beginners to experts.
- [ODSC West Conference- Top 6 Sessions You Must Attend](https://www.iguazio.com/blog/odscwest2021/): ODSC West Reconnect is the place to be for MLOps, data science, and AI. Here are our top 6 recommended sessions for this year's conference.
- [It Worked Fine in Jupyter. Now What? ](https://www.iguazio.com/blog/it-worked-fine-in-jupyter-now-what/): How to use MLRun to quickly deploy applications, and run on Kubernetes without changing code or learning a new technology. 
- [How to Bring Breakthrough Performance and Productivity To AI/ML Projects](https://www.iguazio.com/blog/how-to-bring-breakthrough-performance-and-productivity-to-ai-ml-projects/): Pure Storage and Iguazio empower enterprises to focus on their business applications and not the underlying infrastructure.
- [Building Machine Learning Pipelines with Real-Time Feature Engineering](https://www.iguazio.com/blog/building-real-time-ml-pipelines-with-a-feature-store/): An online feature store enables ML teams to harness real-time data, perform complex calculations in real time and make fast decisions based on fresh data.
- [Implementing Automation and an MLOps Framework for Enterprise-scale ML](https://www.iguazio.com/blog/implementing-automation-and-an-mlops-framework-for-enterprise-scale-ml/): Enterprise AI in production is still immature. Companies are implementing MLOps frameworks to get to production and scale up.
- [Using Automated Model Management for CPG Trade Success](https://www.iguazio.com/blog/using-automated-model-management-for-cpg-trade-success/): AI/ML can help CPG companies execute promotions with greater precision to obtain impactful results and higher ROI.
- [All That Hype: Iguazio Listed in 7 Gartner Hype Cycles for 2021](https://www.iguazio.com/blog/iguazio-listed-in-7-gartner-hype-cycles-for-2021/): We're proud to be mentioned in 7 Gartner Hype Cycles for the second year in a row!
- [Announcing the Winners of the MLOps for Good Hackathon](https://www.iguazio.com/blog/announcing-the-winners-mlops-for-good-hackathon/): We just wrapped up the first-ever MLOps for Good hackathon, and we are so thrilled by the incredible response we’ve gotten from the ML community. 

---


## News

- [Smart Network Maintenance in Africa Powered by Machine Learning](https://www.iguazio.com/news-events/news/smart-network-maintenance-in-africa-powered-by-machine-learning/): Smart Network Maintenance in Africa Powered by Machine Learning . Stay ahead with Iguazio.
- [Iguazio Named Leader in the SPARK Matrix Data Science & Machine Learning Platforms](https://www.iguazio.com/news-events/news/iguazio-recognized-as-a-leader-in-spark-matrix-data-science-and-machine-learning-platforms-q1-2025/): Iguazio Named Leader in the SPARK Matrix Data Science & Machine Learning Platforms . Stay ahead with Iguazio.
- [Iguazio CTO: Successful AI depends on data AND trust](https://www.iguazio.com/news-events/news/iguazio-cto-successful-ai-depends-on-data-and-trust/): Iguazio CTO: Successful AI depends on data AND trust . Stay ahead with Iguazio.
- [Iguazio Named to the Constellation ShortList™ for MLOps Q1 2025](https://www.iguazio.com/news-events/news/iguazio-named-to-the-constellation-shortlist-for-mlops-q1-2025/): Iguazio Named to the Constellation ShortList™ for MLOps Q1 2025 . Stay ahead with Iguazio.
- [An Architect’s Guide to the Top 10 Tools Needed to Build the Modern Data Lake](https://www.iguazio.com/news-events/news/an-architects-guide-to-the-top-10-tools-needed-to-build-the-modern-data-lake/): An Architect’s Guide to the Top 10 Tools Needed to Build the Modern Data Lake . Stay ahead with Iguazio.
- [McKinsey offering aims to bridge the gap from AI prototypes to production](https://www.iguazio.com/news-events/news/mckinsey-offering-aims-to-bridge-the-gap-from-ai-prototypes-to-production-2/): McKinsey offering aims to bridge the gap from AI prototypes to production . Stay ahead with Iguazio.
- [2024 Top Performer MLOps Platform](https://www.iguazio.com/news-events/news/2024-top-performer-mlops-platform/): 2024 Top Performer MLOps Platform . Stay ahead with Iguazio.
- [AiThority Interview with Asaf Somekh, Co-Founder & CEO of Iguazio (acquired by McKinsey)](https://www.iguazio.com/news-events/news/aithority-interview-with-asaf-somekh-co-founder-ceo-of-iguazio-acquired-by-mckinsey/): AiThority Interview with Asaf Somekh, Co-Founder & CEO of Iguazio (acquired by McKinsey) . Stay ahead with Iguazio.
- [5 Best End-to-End Open Source MLOps Tools](https://www.iguazio.com/news-events/news/5-best-end-to-end-open-source-mlops-tools/): 5 Best End-to-End Open Source MLOps Tools . Stay ahead with Iguazio.
- [The Architect’s Guide to the GenAI Tech Stack — 10 Tools](https://www.iguazio.com/news-events/news/mckinsey-offering-aims-to-bridge-the-gap-from-ai-prototypes-to-production/): The Architect’s Guide to the GenAI Tech Stack — 10 Tools . Stay ahead with Iguazio.
- [McKinsey accelerates gen AI value creation with Iguazio](https://www.iguazio.com/news-events/news/mckinsey-accelerates-gen-ai-value-creation-with-iguazio/): McKinsey accelerates gen AI value creation with Iguazio . Stay ahead with Iguazio.
- [Iguazio named to Constellation's ShortList™ MLOps – Feb 2024](https://www.iguazio.com/news-events/news/iguazio-named-to-constellations-shortlist-mlops-feb-2024/): Iguazio named to Constellation's ShortList™ MLOps – Feb 2024 . Stay ahead with Iguazio.
- [McKinsey & Company receives the MongoDB 2024 Transformation Partner Award for its work with Iguazio](https://www.iguazio.com/news-events/news/mckinsey-company-receives-the-mongodb-2024-transformation-partner-award-for-its-work-with-iguazio/): McKinsey & Company receives the MongoDB 2024 Transformation Partner Award for its work with Iguazio . Stay ahead with Iguazio.
- [Iguazio named to the CB Insights LLMOps (Large Language Model Operations) Market Map](https://www.iguazio.com/news-events/news/iguazio-named-to-the-cb-insights-llmops-large-language-model-operations-market-map/): Iguazio named to the CB Insights LLMOps (Large Language Model Operations) Market Map . Stay ahead with Iguazio.
- [The 21 Best Artificial Intelligence Platforms Of 2024](https://www.iguazio.com/news-events/news/the-21-best-artificial-intelligence-platforms-of-2024/): The 21 Best Artificial Intelligence Platforms Of 2024 . Stay ahead with Iguazio.
- [Iguazio Listed in Constellation's ShortList™ MLOps](https://www.iguazio.com/news-events/news/iguazio-listed-in-constellations-shortlist-mlops/): Iguazio Listed in Constellation's ShortList™ MLOps . Stay ahead with Iguazio.
- [The (un)real world of Generative AI](https://www.iguazio.com/news-events/news/the-unreal-world-of-generative-ai/): The (un)real world of Generative AI . Stay ahead with Iguazio.
- [Musk, AI, And 'Civilizational Destruction: Prophecy or Product Launch?](https://www.iguazio.com/news-events/news/musk-ai-and-civilizational-destruction-prophecy-or-product-launch/): Musk, AI, And 'Civilizational Destruction: Prophecy or Product Launch? . Stay ahead with Iguazio.
- [McKinsey acquires Iguazio, a leader in AI and machine-learning technology](https://www.iguazio.com/news-events/news/iguazio-acquired-by-mckinsey/): McKinsey acquires Iguazio, a leader in AI and machine-learning technology . Stay ahead with Iguazio.
- [Iguazio Named a Major Player in the IDC MLOps MarketScape 2022](https://www.iguazio.com/news-events/news/iguazio-named-a-major-player-in-the-idc-mlops-marketscape-2022/): With Iguazio, Sense powers a wide range of AI products aimed at increasing the efficiency and scalability of their AI operations.
- [Iguazio Named a Leader and Outperformer In GigaOm Radar for MLOps 2022](https://www.iguazio.com/news-events/news/iguazio-named-a-leader-and-outperformer-in-gigaom-radar-for-mlops-2022/): With Iguazio, Sense powers a wide range of AI products aimed at increasing the efficiency and scalability of their AI operations.
- [Iguazio Named in 8 Gartner Hype Cycles for 2022](https://www.iguazio.com/news-events/news/iguazio-named-in-8-gartner-hype-cycles-for-2022/): With Iguazio, Sense powers a wide range of AI products aimed at increasing the efficiency and scalability of their AI operations.
- [Sense Selects Iguazio for AI Chatbot Automation ](https://www.iguazio.com/news-events/news/sense-selects-iguazio-for-ai-chatbot-automation/): With Iguazio, Sense powers a wide range of AI products aimed at increasing the efficiency and scalability of their AI operations.
- [Iguazio Partners with Snowflake to Automate and Accelerate MLOps](https://www.iguazio.com/news-events/news/iguazio-partners-with-snowflake-to-automate-and-accelerate-mlops/): Iguazio Partners with Snowflake to Automate and Accelerate MLOps . Stay ahead with Iguazio.
- [Gartner 2022 Market Guide for DSML Engineering Platforms](https://www.iguazio.com/news-events/news/gartner-2022-market-guide-for-dsml-engineering-platforms/): Gartner 2022 Market Guide for DSML Engineering Platforms . Stay ahead with Iguazio.
- [Iguazio named in The Coolest Data Science And Machine Learning Tool Companies Of The 2022 Big Data 100](https://www.iguazio.com/news-events/news/the-coolest-data-science-and-machine-learning-tool-companies-of-the-2022-big-data-100/): Iguazio named in The Coolest Data Science And Machine Learning Tool Companies Of The 2022 Big Data 100 . Stay ahead with Iguazio.
- [Iguazio named in Forrester's Now Tech: AI/ML Platforms, Q1 2022](https://www.iguazio.com/news-events/news/iguazio-named-in-forresters-now-tech-ai-ml-platforms-q1-2022/): Iguazio named in Forrester's Now Tech: AI/ML Platforms, Q1 2022 . Stay ahead with Iguazio.
- [Iguazio Named a Fast-Moving Leader by Gigaom in the Radar for MLOps Report](https://www.iguazio.com/news-events/news/iguazio-named-a-fast-moving-leader-by-gigaom-in-the-radar-for-mlops-report-2/): Iguazio Named a Fast-Moving Leader by Gigaom in the Radar for MLOps Report . Stay ahead with Iguazio.
- [All That Hype: Iguazio Listed in 5 Gartner Hype Cycles for 2021](https://www.iguazio.com/news-events/news/all-that-hype-iguazio-listed-in-5-gartner-hype-cycles-for-2021/): All That Hype: Iguazio Listed in 5 Gartner Hype Cycles for 2021 . Stay ahead with Iguazio.
- [Iguazio Partners with Pure Storage to Operationalize AI for Enterprises](https://www.iguazio.com/news-events/news/iguazio-announces-mlops-for-good-virtual-hackathon-2/): Iguazio Partners with Pure Storage to Operationalize AI for Enterprises . Stay ahead with Iguazio.
- [Iguazio MLOps Platform Launches in AWS Marketplace](https://www.iguazio.com/news-events/news/iguazio-mlops-platform-launches-in-aws-marketplace/): Iguazio MLOps Platform Launches in AWS Marketplace . Stay ahead with Iguazio.
- [MLOps: The Latest Shift in the AI Market in Israel](https://www.iguazio.com/news-events/news/mlops-the-latest-shift-in-the-ai-market-in-israel/): MLOps: The Latest Shift in the AI Market in Israel . Stay ahead with Iguazio.
- [Iguazio Announces ‘MLOps for Good’ Virtual Hackathon](https://www.iguazio.com/news-events/news/iguazio-announces-mlops-for-good-virtual-hackathon/): Iguazio Announces ‘MLOps for Good’ Virtual Hackathon . Stay ahead with Iguazio.
- [Boston Limited and Iguazio Partner to Operationalize AI for the Enterprise](https://www.iguazio.com/news-events/news/boston-limited-and-iguazio-partner-to-operationalize-ai-for-the-enterprise/): Boston Limited and Iguazio Partner to Operationalize AI for the Enterprise . Stay ahead with Iguazio.
- [Iguazio Named A Fast Moving Leader by GigaOm in the ‘Radar for MLOps’ Report](https://www.iguazio.com/news-events/news/iguazio-named-a-fast-moving-leader-by-gigaom-in-the-radar-for-mlops-report/): Iguazio Named A Fast Moving Leader by GigaOm in the ‘Radar for MLOps’ Report . Stay ahead with Iguazio.
- [The Coolest Data Science And Machine Learning Tool Companies Of The 2021 Big Data 100](https://www.iguazio.com/news-events/news/the-coolest-data-science-and-machine-learning-tool-companies-of-the-2021-big-data-100/): The Coolest Data Science And Machine Learning Tool Companies Of The 2021 Big Data 100 . Stay ahead with Iguazio.
- [Iguazio Named Leader and Fast Mover in GigaOm Radar for Evaluating Machine Learning Operations (MLOps)](https://www.iguazio.com/news-events/news/gigaom-radar-for-evaluating-machine-learning-operations-mlops/): Iguazio Named Leader and Fast Mover in GigaOm Radar for Evaluating Machine Learning Operations (MLOps) . Stay ahead with Iguazio.
- [The Next-Level of Operationalizing Machine Learning: Real-time Data Streaming into Data Science Environments](https://www.iguazio.com/news-events/news/the-next-level-of-operationalizing-machine-learning-real-time-data-streaming-into-data-science-environments/): The Next-Level of Operationalizing Machine Learning: Real-time Data Streaming into Data Science Environments . Stay ahead with Iguazio.
- [Iguazio Receives an Honorable Mention in the 2021 Gartner Magic Quadrant for Data Science and Machine Learning Platforms](https://www.iguazio.com/news-events/news/iguazio-receives-an-honorable-mention-in-the-2021-magic-quadrant-for-data-science-and-machine-learning-platforms/): Iguazio Receives an Honorable Mention in the 2021 Gartner Magic Quadrant for Data Science and Machine Learning Platforms . Stay ahead with Iguazio.
- [The AI Infrastructure Alliance Launches With 25 Members to Create the Canonical Stack for Artificial Intelligence Projects](https://www.iguazio.com/news-events/news/the-ai-infrastructure-alliance-launches-with-25-members-to-create-the-canonical-stack-for-artificial-intelligence-projects/): The AI Infrastructure Alliance Launches With 25 Members to Create the Canonical Stack for Artificial Intelligence Projects . Stay ahead with Iguazio.
- [Git-based CI/CD for Machine Learning and MLOps](https://www.iguazio.com/news-events/news/git-based-ci-cd-for-machine-learning-and-mlops/): Git-based CI/CD for Machine Learning and MLOps . Stay ahead with Iguazio.
- [Iguazio Signs Strategic Agreement with Sheba Medical Center for Real-Time Covid-19 Treatment](https://www.iguazio.com/news-events/news/iguazio-signs-strategic-agreement-with-sheba-medical-center-for-real-time-covid-19-treatment/): Iguazio Signs Strategic Agreement with Sheba Medical Center for Real-Time Covid-19 Treatment . Stay ahead with Iguazio.
- [Sheba, Iguazio to Develop Real-Time AI to Optimise Patient Care](https://www.iguazio.com/news-events/news/sheba-iguazio-to-develop-real-time-ai-to-optimise-patient-care/): Sheba, Iguazio to Develop Real-Time AI to Optimise Patient Care . Stay ahead with Iguazio.
- [Iguazio Launches the First Integrated Feature Store Within its Data Science Platform](https://www.iguazio.com/news-events/news/iguazio-launches-the-first-integrated-feature-store-within-its-data-science-platform/): Iguazio Launches the First Integrated Feature Store Within its Data Science Platform . Stay ahead with Iguazio.
- [The 12 Coolest Machine-Learning Startups Of 2020](https://www.iguazio.com/news-events/news/the-12-coolest-machine-learning-startups-of-2020/): The 12 Coolest Machine-Learning Startups Of 2020 . Stay ahead with Iguazio.
- [An AI Engineer Walks Into A Data Shop...](https://www.iguazio.com/news-events/news/an-ai-engineer-walks-into-a-data-shop/): An AI Engineer Walks Into A Data Shop... . Stay ahead with Iguazio.
- [Why You Need to Start Thinking About MLOps](https://www.iguazio.com/news-events/news/why-you-need-to-start-thinking-about-mlops/): Why You Need to Start Thinking About MLOps . Stay ahead with Iguazio.
- [SFL Scientific And Iguazio Partner To Speed Up Custom AI Development For Fortune 1000 Companies](https://www.iguazio.com/news-events/news/sfl-scientific-and-iguazio-partner-to-speed-up-custom-ai-development-for-fortune-1000-companies/): SFL Scientific And Iguazio Partner To Speed Up Custom AI Development For Fortune 1000 Companies . Stay ahead with Iguazio.
- [SFL Scientific and Iguazio Partner to Speed Up Custom AI Development for Fortune 1000 companies](https://www.iguazio.com/news-events/news/sfl-scientific-and-iguazio-partner-to-speed-up-custom-ai-development-for-fortune-1000-companies-2/): SFL Scientific and Iguazio Partner to Speed Up Custom AI Development for Fortune 1000 companies . Stay ahead with Iguazio.
- [NetApp Deploys Iguazio to Run AI-Driven Digital Advisor on Active IQ](https://www.iguazio.com/news-events/news/netapp-deploys-iguazio-to-run-ai-driven-digital-advisor-on-active-iq/): NetApp Deploys Iguazio to Run AI-Driven Digital Advisor on Active IQ . Stay ahead with Iguazio.
- [Iguazio Becomes Certified for NVIDIA DGX-Ready Software Program](https://www.iguazio.com/news-events/news/iguazio-becomes-certified-for-nvidia-dgx-ready-software-program/): Iguazio Becomes Certified for NVIDIA DGX-Ready Software Program . Stay ahead with Iguazio.
- [Iguazio & NetApp Partner to Accelerate Deployment of AI](https://www.iguazio.com/news-events/news/iguazio-netapp-partner-to-accelerate-deployment-of-ai/): Iguazio & NetApp Partner to Accelerate Deployment of AI . Stay ahead with Iguazio.
- [NetApp, Iguazio Build Joint Tech To Accelerate AI Deployments](https://www.iguazio.com/news-events/news/netapp-iguazio-build-joint-tech-to-accelerate-ai-deployments/): NetApp, Iguazio Build Joint Tech To Accelerate AI Deployments . Stay ahead with Iguazio.
- [The Coolest Data Science And Machine Learning Tool Companies Of The 2020 Big Data 100](https://www.iguazio.com/news-events/news/the-coolest-data-science-and-machine-learning-tool-companies-of-the-2020-big-data-100/): The Coolest Data Science And Machine Learning Tool Companies Of The 2020 Big Data 100 . Stay ahead with Iguazio.
- [Enabling end-to-end machine learning workflows with Iguazio](https://www.iguazio.com/news-events/news/enabling-end-to-end-machine-learning-workflows-with-iguazio-2/): Streamline machine learning processes using Iguazio's comprehensive end-to-end workflow solutions for seamless efficiency and innovation.
- [Iguazio Receives an Honorable Mention in the Gartner MQ for Data Science and ML Platforms](https://www.iguazio.com/news-events/news/gartner-gives-iguazio-an-honorable-mention-in-its-2020-magic-quadrant-for-data-science-and-machine-learning-platforms/): Iguazio Receives an Honorable Mention in the Gartner MQ for Data Science and ML Platforms. Read more here.
- [Iguazio raises $24 million for AI development and management](https://www.iguazio.com/news-events/news/iguazio-raises-24-million-for-ai-development-and-management-tools/): Iguazio raises $24 million for AI development and management . Stay ahead with Iguazio.
- [Dell Technologies Introduces New Solutions with Iguazio](https://www.iguazio.com/news-events/news/dell-technologies-introduces-new-solutions-to-advance-high-performance-computing-and-ai-innovation/): Dell Technologies Introduces New Solutions with Iguazio . Stay ahead with Iguazio.
- [PICSIX Partners with Iguazio](https://www.iguazio.com/news-events/news/picsix-launches-an-investigative-intelligence-platform-powered-by-iguazios-real-time-data-science-platform/): PICSIX Partners with Iguazio . Stay ahead with Iguazio.
- [The Rise of MLOps: What We Can All Learn from DevOps](https://www.iguazio.com/news-events/news/the-rise-of-mlops-what-we-can-all-learn-from-devops/): The Rise of MLOps: What We Can All Learn from DevOps . Stay ahead with Iguazio.
- [Bringing AI and Machine Learning to the Masses](https://www.iguazio.com/news-events/news/bringing-ai-and-machine-learning-to-the-masses/): Bringing AI and Machine Learning to the Masses . Stay ahead with Iguazio.
- [Takeaway from MLOps NYC: Open Source Frameworks Need TLC](https://www.iguazio.com/news-events/news/takeaway-from-mlops-nyc-open-source-frameworks-need-tlc/): Takeaway from MLOps NYC: Open Source Frameworks Need TLC . Stay ahead with Iguazio.
- [Top 10 IoT Startups Of 2019](https://www.iguazio.com/news-events/news/top-10-iot-startups-of-2019/): Top 10 IoT Startups Of 2019 . Stay ahead with Iguazio.
- [Hitting the Reset Button on Hadoop](https://www.iguazio.com/news-events/news/hitting-the-reset-button-on-hadoop/): Hitting the Reset Button on Hadoop . Stay ahead with Iguazio.
- [CEO Q&A: Modern Platforms for Data Science](https://www.iguazio.com/news-events/news/ceo-qa-modern-platforms-for-data-science/): CEO Q&A: Modern Platforms for Data Science . Stay ahead with Iguazio.
- [Iguazio Brings Its Data Science Platform to Azure and Azure Stack](https://www.iguazio.com/news-events/news/iguazio-brings-its-data-science-platform-to-azure-and-azure-stack/): Iguazio Brings Its Data Science Platform to Azure and Azure Stack . Stay ahead with Iguazio.
- [Samsung SDS Backs Data Company Iguazio](https://www.iguazio.com/news-events/news/samsung-sds-backs-data-company-iguazio/): Samsung SDS Backs Data Company Iguazio . Stay ahead with Iguazio.
- [Q&A with Iguazio: on Data Science, Data Analytics, and Serverless](https://www.iguazio.com/news-events/news/qa-with-iguazio-on-data-science-data-analytics-and-serverless/): Q&A with Iguazio: on Data Science, Data Analytics, and Serverless . Stay ahead with Iguazio.
- [How Serverless Platforms Could Power an Event-Driven AI Pipeline](https://www.iguazio.com/news-events/news/how-serverless-platforms-could-power-an-event-driven-ai-pipeline/): How Serverless Platforms Could Power an Event-Driven AI Pipeline . Stay ahead with Iguazio.
- [Removing Data Blockage at the Edge](https://www.iguazio.com/news-events/news/removing-data-blockage-at-the-edge/): Removing Data Blockage at the Edge . Stay ahead with Iguazio.
- [The 10 Coolest New Open-Source Technologies And Tools Of 2018](https://www.iguazio.com/news-events/news/the-10-coolest-new-open-source-technologies-and-tools-of-2018/): The 10 Coolest New Open-Source Technologies And Tools Of 2018 . Stay ahead with Iguazio.
- [Google Cloud collaborating with Iguazio to enable real-time AI across the cloud and intelligent edge](https://www.iguazio.com/news-events/news/google-cloud-collaborating-with-iguazio-to-enable-real-time-ai-across-the-cloud-and-intelligent-edge/): Google Cloud collaborating with Iguazio to enable real-time AI across the cloud and intelligent edge . Stay ahead with Iguazio.
- [Don’t get cloudwashed: The case for cloud on-prem in hybrid computing](https://www.iguazio.com/news-events/news/dont-get-cloudwashed-the-case-for-cloud-on-prem-in-hybrid-computing/): Don’t get cloudwashed: The case for cloud on-prem in hybrid computing . Stay ahead with Iguazio.
- [Even in the cloud, banking is tied to legacy tech](https://www.iguazio.com/news-events/news/even-in-the-cloud-banking-is-tied-to-legacy-tech/): Even in the cloud, banking is tied to legacy tech . Stay ahead with Iguazio.
- [Other Vendors to Consider for Operational DBMSs](https://www.iguazio.com/news-events/news/other-vendors-to-consider-for-operational-dbmss/): Other Vendors to Consider for Operational DBMSs . Stay ahead with Iguazio.
- [Iguazio’s New Nuclio Release Enables Serverless Agility for Enterprises Deploying Real-time Intelligent Applications](https://www.iguazio.com/news-events/news/iguazios-new-nuclio-release-enables-serverless-agility-for-enterprises-deploying-real-time-intelligent-applications/): Iguazio’s New Nuclio Release Enables Serverless Agility for Enterprises Deploying Real-time Intelligent Applications . Stay ahead with Iguazio.
- [SD Times Blog: Getting a serverless reality check](https://www.iguazio.com/news-events/news/sd-times-blog-getting-a-serverless-reality-check/): SD Times Blog: Getting a serverless reality check . Stay ahead with Iguazio.
- [Iguazio revamps its Nuclio serverless computing platform](https://www.iguazio.com/news-events/news/iguazio-revamps-its-nuclio-serverless-computing-platform/): Iguazio revamps its Nuclio serverless computing platform . Stay ahead with Iguazio.
- [SD Times news digest: Iguazio’s Nuclio release, Kotlin 1.3, and reCAPTCHA v3](https://www.iguazio.com/news-events/news/sd-times-news-digest-iguazios-nuclio-release-kotlin-1-3-and-recaptcha-v3/): SD Times news digest: Iguazio’s Nuclio release, Kotlin 1.3, and reCAPTCHA v3 . Stay ahead with Iguazio.
- [Iguazio Selected for CNBC's Upstart 100 List of Promising Startups](https://www.iguazio.com/news-events/news/iguazio-selected-for-cnbcs-upstart-100-list-of-promising-startups/): Iguazio Selected for CNBC's Upstart 100 List of Promising Startups . Stay ahead with Iguazio.
- [Instead of sending data to the cloud, why not send the cloud to the edge?](https://www.iguazio.com/news-events/news/instead-of-sending-data-to-the-cloud-why-not-send-the-cloud-to-the-edge/): Instead of sending data to the cloud, why not send the cloud to the edge? . Stay ahead with Iguazio.
- [Is cloud native starting to kill Hadoop? This CTO says yes](https://www.iguazio.com/news-events/news/is-cloud-native-starting-to-kill-hadoop-this-cto-says-yes/): Is cloud native starting to kill Hadoop? This CTO says yes . Stay ahead with Iguazio.
- [This startup thinks it knows how to speed up real-time analytics on tons of data](https://www.iguazio.com/news-events/news/this-startup-thinks-it-knows-how-to-speed-up-real-time-analytics-on-tons-of-data/): This startup thinks it knows how to speed up real-time analytics on tons of data . Stay ahead with Iguazio.
- [Equinix and Iguazio partner to drive smart mobility](https://www.iguazio.com/news-events/news/equinix-and-iguazio-partner-to-drive-smart-mobility/): Equinix and Iguazio partner to drive smart mobility . Stay ahead with Iguazio.
- [Enabling smart transportation in today's interconnected world](https://www.iguazio.com/news-events/news/enabling-smart-transportation-in-todays-interconnected-world/): Enabling smart transportation in today's interconnected world . Stay ahead with Iguazio.
- [Iguazio’s Nuclio Serverless Software Aims to Outrun AWS](https://www.iguazio.com/news-events/news/iguazios-nuclio-serverless-software-aims-to-outrun-aws/): Iguazio’s Nuclio Serverless Software Aims to Outrun AWS . Stay ahead with Iguazio.
- [The Car's Eyes and Ears at TC TLV 2018](https://www.iguazio.com/news-events/news/the-cars-eyes-and-ears-at-tc-tlv-2018/): The Car's Eyes and Ears at TC TLV 2018 . Stay ahead with Iguazio.
- [Enterprise startups in Israel worth getting to know](https://www.iguazio.com/news-events/news/enterprise-startups-in-israel-worth-getting-to-know/): Enterprise startups in Israel worth getting to know . Stay ahead with Iguazio.
- [Bigger than Linux: The rise of cloud native](https://www.iguazio.com/news-events/news/bigger-than-linux-the-rise-of-cloud-native/): Bigger than Linux: The rise of cloud native . Stay ahead with Iguazio.
- [2018 Big Data 100: 45 Coolest Data Management And Integration Vendors](https://www.iguazio.com/news-events/news/2018-big-data-100-45-coolest-data-management-and-integration-vendors/): 2018 Big Data 100: 45 Coolest Data Management And Integration Vendors . Stay ahead with Iguazio.
- [Serverless computing takes a big step into the multicloud world](https://www.iguazio.com/news-events/news/serverless-computing-takes-a-big-step-into-the-multicloud-world/): Serverless computing takes a big step into the multicloud world . Stay ahead with Iguazio.
- [Serverless framework Nuclio released for enterprise customers](https://www.iguazio.com/news-events/news/serverless-framework-nuclio-released-for-enterprise-customers/): Serverless framework Nuclio released for enterprise customers . Stay ahead with Iguazio.
- [This Israeli Startup Partners With Amazon -- But Could Compete With AWS In $8B Edge Cloud Market](https://www.iguazio.com/news-events/news/this-israeli-startup-partners-with-amazon-but-could-compete-with-aws-in-8b-edge-cloud-market/): This Israeli Startup Partners With Amazon -- But Could Compete With AWS In $8B Edge Cloud Market . Stay ahead with Iguazio.
- [Can Open-Source Serverless Be Simpler than AWS Lambda?](https://www.iguazio.com/news-events/news/can-open-source-serverless-be-simpler-than-aws-lambda/): Can Open-Source Serverless Be Simpler than AWS Lambda? . Stay ahead with Iguazio.
- [Asaf Somekh, Founder & CEO at iguazio, talks delivering dreams](https://www.iguazio.com/news-events/news/asaf-somekh-founder-ceo-at-iguazio-talks-delivering-dreams/): Asaf Somekh, Founder & CEO at iguazio, talks delivering dreams . Stay ahead with Iguazio.
- [To reach its full promise, big data must begin with a clean slate](https://www.iguazio.com/news-events/news/to-reach-its-full-promise-big-data-must-begin-with-a-clean-slate/): To reach its full promise, big data must begin with a clean slate . Stay ahead with Iguazio.
- [Big Data Startup iguazio Debuts Its First Channel Program, Seeks VARs and SIs with Vertical Industry Expertise](https://www.iguazio.com/news-events/news/big-data-startup-iguazio-debuts-its-first-channel-program-seeks-vars-and-sis-with-vertical-industry-expertise/): Big Data Startup iguazio Debuts Its First Channel Program, Seeks VARs and SIs with Vertical Industry Expertise . Stay ahead with Iguazio.
- [Asia Pacific expansion on the horizon for iguazio with opening of Singapore headquarters](https://www.iguazio.com/news-events/news/asia-pacific-expansion-on-the-horizon-for-iguazio-with-opening-of-singapore-headquarters/): Asia Pacific expansion on the horizon for iguazio with opening of Singapore headquarters . Stay ahead with Iguazio.
- [iguazio Boldly Taunts AWS’ Lambda with nuclio Serverless Platform](https://www.iguazio.com/news-events/news/iguazio-boldly-taunts-aws-lambda-with-nuclio-serverless-platform/): iguazio Boldly Taunts AWS’ Lambda with nuclio Serverless Platform . Stay ahead with Iguazio.
- [Surge pricing: How it works and how to avoid it](https://www.iguazio.com/news-events/news/surge-pricing-how-it-works-and-how-to-avoid-it/): Surge pricing: How it works and how to avoid it . Stay ahead with Iguazio.
- [Serverless Framework for Real-Time Apps Emerges](https://www.iguazio.com/news-events/news/serverless-framework-for-real-time-apps-emerges/): Serverless Framework for Real-Time Apps Emerges . Stay ahead with Iguazio.

---


## Events

- [MLOps Live #39 - Orchestrating Multi-Agent Workflows with MCP & A2A: The Future of Scalable AI Systems](https://www.iguazio.com/events/mlops-live-39/): Join Iguazio CTO and Google Cloud expert as they explore cutting-edge tools enabling scalable AI workflows.
- [MLOps Live #37 Building Agent Co-pilots for Proactive Call Centers](https://www.iguazio.com/events/mlops-live-37-building-agent-co-pilots-for-proactive-call-centers/): This webinar explores how GenAI-powered agent co-pilots can identify missed opportunities, deliver real-time personalized insights, and help agents close deals...
- [MLOps Live #36 How to Manage Thousands of Real-Time Models in Production](https://www.iguazio.com/events/mlops-live-36-how-to-manage-thousands-of-real-time-models-in-production/): We were thrilled to welcome back Seagate to share the progress they made with their ML pipeline & MLOps automation...
- [MWC 2025](https://www.iguazio.com/events/mwc-2025/): Our AI and gen AI experts will be on the ground at MWC 2025 and we'd love to meet you...
- [MLOps Live #38 - LLM Evaluation and Testing for Reliable AI Apps](https://www.iguazio.com/events/mlops-live-35-beyond-the-hype-gen-ai-trends-and-scaling-strategies-for-2025/): Evidently AI is joining us to discuss key strategies for evaluating and optimizing LLMs. Learn how to improve performance, mitigate...
- [MLOps Live #34 - Agentic AI Frameworks: Bridging Foundation Models and Business Impact](https://www.iguazio.com/events/mlops-live-34-agentic-ai-frameworks-bridging-foundation-models-and-business-impact/): Discover how conversational AI transforms customer experiences - watch to learn practical tips for building impactful AI agents.
- [MLOps Live #33 - Deploying Gen AI in Production with NVIDIA NIM & MLRun](https://www.iguazio.com/events/mlops-live-33-deploying-gen-ai-in-production-with-nvidia-nim-mlrun/): In this webinar, we explored how to successfully deploy your Gen AI applications while mitigating these challenges, using NVIDIA NIM...
- [MLOps Live #32 - Gen AI for Marketing - From Hype to Implementation](https://www.iguazio.com/events/mlops-live-32-gen-ai-for-marketing-from-hype-to-implementation/): In this session we were joined by a Modern Marketing Capabilities Leader at McKinsey, to delve into how data scientists...
- [MLOps Live #31 - Building Scalable Customer-Facing Gen AI Applications Effectively & Responsibly](https://www.iguazio.com/events/mlops-live-31/): In this session, we shared the Iguazio & MongoDB one-stop-shop solution for building gen AI applications that scale effectively and...
- [MLOPs Live #30 - Implementing Gen AI in Highly Regulated Environments](https://www.iguazio.com/events/mlops-live-30-implementing-gen-ai-in-highly-regulated-environmentsmlops-live-30/): Our CEO & CTO discussed implementing gen AI in highly regulated industries and innovative ways to mitigate them. We shared...
- [MLOps Live #29 - Transforming Enterprise Operations with Gen AI](https://www.iguazio.com/events/mlops-live-29-transforming-enterprise-operations-with-gen-ai/): In our webinar with McKinsey, we dove into the transformative impact of gen AI on enterprise operations, spotlighting advancements across...
- [MLOps Live #28 - Improving LLM Accuracy & Performance](https://www.iguazio.com/events/mlops-live-28-improving-llm-accuracy-performance/): Watch our session with Databricks to hear advice on improving the accuracy & performance of LLMs while mitigating challenges like...
- [MLOps Live #27 -  LLM Validation & Evaluation](https://www.iguazio.com/events/mlops-live-27/): In this webinar we demonstrated how to effectively validate and evaluate your LLM, and dove into the pipeline to show...
- [MLOps Live #26 - Implementing a Gen AI Smart Call Center Analysis App](https://www.iguazio.com/events/mlops-live-26-implementing-a-gen-ai-smart-call-center-analysis-app/): In this session, Oana Cheta Partner & Lead Gen AI Service Ops NA at McKinsey & Company joined Yaron Haviv...
- [MLOps Live #25 - GenAI for Financial Services](https://www.iguazio.com/events/mlops-live-25-genai-for-financial-servicesmlops-live-25/): In this session we discussed how leading global Financial Service companies are using GenAI today to serve their clients better,...
- [MLOps Live #24 - How to Build an Automated AI ChatBot](https://www.iguazio.com/events/mlops-live-23-mlops-for-generative-ai/): In this session, Gennaro, Director of Data Science at Sense, share how he and his team have built and perfected...
- [ODSC West 2023](https://www.iguazio.com/events/odsc-west-2023/): Watch Yaron Haviv, our Co-Founder and CTO, present the MLOps Keynote at ODSC West 2023.
- [AI at Scale](https://www.iguazio.com/events/ai-at-scale/): We are excited to be speaking at AIIA's upcoming event! Join us as we dive deep into scaling AI, including...
- [NVIDIA GTC](https://www.iguazio.com/events/nvidia-gtc-2/): We are excited to be participating in NVIDIA GTC 2023. Yaron Haviv, Iguazio's Co-Founder and CTO, will be presenting as...
- [Future of AI 2022](https://www.iguazio.com/events/future-of-ai-2022/): We are excited to be sponsoring the Future of AI Conference 2022. Future of AI brings together the leading Israeli...
- [MLOps Live #22 How Seagate Handles Data Engineering at Scale](https://www.iguazio.com/events/mlops-live-21-hcis-journey-to-mlops-efficiency/): Join Vamsi and Yaron for a fascinating deep dive into Seagate's journey to MLOps efficiency. Hear Seagate’s story, watch a...
- [MLOps Live #20: How to Easily Deploy Your Hugging Face Model to Production at Scale](https://www.iguazio.com/events/mlops-live-20-how-to-easily-deploy-your-hugging-face-model-to-production-at-scale/): Join us for this technical session with Hugging Face, and learn from the experts!
- [AIIA: Data-Centric AI Summit](https://www.iguazio.com/events/aiia-data-centric-ai-summit/): In our session, we will demonstrate how to use Iguazio & Snowflake to create a simple, seamless, and automated path...
- [TMLS MLOps World: Conference on Machine Learning in Production 2022](https://www.iguazio.com/events/tmls-mlops-world-conference-on-machine-learning-in-production-2022/): We are proud to be Gold Sponsors of the 3rd annual MLOps World Summit Conference and Expo. MLOps World is...
- [Snowflake Summit](https://www.iguazio.com/events/snowflake-summit/): We are proud to be sponsoring the 2022 Snowflake Summit. The summit gathers top technical, data, and business experts to...
- [ODSC Europe](https://www.iguazio.com/events/odsc-europe-london/): ODSC Europe combines immersive in-person sessions and hands-on training with innovative virtual ones. We hope to see you there!
- [MLOps: Machine Learning in Production / New York City Summit](https://www.iguazio.com/events/mlops-machine-learning-in-production-new-york-city-summit/): We are excited to be a featured speaker for NYC's virtual ML in Production conference. This event gathers industry leaders...
- [ODSC East](https://www.iguazio.com/events/odsc-east-boston/): We are proud sponsors of the ODSC East Conference, and the keynote speaker of the Data Engineering and MLOps track.
- [Understanding Fraud Prediction for Banking & Finance Sector through Iguazio & Royal Cyber](https://www.iguazio.com/events/webinar-understanding-fraud-prediction-for-banking-finance-sector-through-iguazio-royal-cyber/): The webinar will focus on how banking and financial institutions can use cutting-edge machine learning and AI tools and technologies...
- [MDLI Ops Conference 2023](https://www.iguazio.com/events/mdli-ops-conference-2022/): We are proud to be speaking and sponsoring MDLI Ops 2023. This yearly conference brings together experts in the field...
- [ODSC Webinar: Git Based CI/CD for ML](https://www.iguazio.com/events/git-based-ci-cd-for-ml-a-complimentary-odsc-webinar/): In this live webinar we examine naive ML workflows and explore why they break down, showcase system design principles that...
- [Session #17: Scaling NLP Pipelines at IHS Markit](https://www.iguazio.com/events/session-17-scaling-nlp-pipelines-at-ihs-markit/): MLOps LIVE is back and better than ever! The data science team at IHS Markit will be sharing practical advice...
- [GTC](https://www.iguazio.com/events/gtc/): We are excited to be participating in GTC 2021. Join us as we accompany our partners on a panel: Mastering...
- [NetApp Insight 2021](https://www.iguazio.com/events/netapp-insight-201/): Pop by our booth to say hi!
- [ODSC West](https://www.iguazio.com/events/odsc-west/): We are excited to be returning to ODSC! Mark your calendars for the joint Track Keynote with IHS Markit: "MLOps...
- [Iguazio MLOps Platform Launches in AWS Marketplace](https://www.iguazio.com/events/iguazio-mlops-platform-launches-in-aws-marketplace/): Iguazio announced its availability in the AWS Marketplace, a digital catalog with thousands of software listings from independent software vendors...
- [Kubecon North America 2021](https://www.iguazio.com/events/kubecon-north-america-2021/): Stay tuned for more information.
- [MLOps: Machine Learning in Production New York City](https://www.iguazio.com/events/mlops-machine-learning-in-production-new-york-city/): We are proud to be sponsoring this years MLOps NYC Conference dedicated to propogating clearer understanding of best practices, methodologies,...
- ["Building a Real-Time ML Pipeline with a Feature Store" - MLOps Live Webinar #16](https://www.iguazio.com/events/mlops-live-webinar-series-session-16/): Save your spot for our upcoming MLOps webinar and learn about challenges associated with online feature engineering, how feature stores...
- [#MLOpsforGood Award Ceremony](https://www.iguazio.com/events/mlopsforgood-award-ceremony/): After six excellent weeks of hands-on hacking, 300 participants, and over 30 awesome projects, we invite you to the live...
- [MLOps in Finance Summit](https://www.iguazio.com/events/mlops-in-finance-summit/): Catch us in the upcoming MLOps in Finance Summit! Yaron Haviv, Iguazio's CTO, will be presenting on how to build...
- [MLOps for Good Hackathon](https://www.iguazio.com/events/mlops-for-good-hackathon/): Join us for the first-ever MLRun hackathon, to help bring data science to production for social Good. Iguazio will be...
- [MLOps World: Machine Learning in Production Conference](https://www.iguazio.com/events/mlops-world-machine-learning-in-production-conference/): Join Iguazio at the 2nd Annual MLOps World Conference on Machine Learning in Production. We invite you to attend our...
- [ODSC EUROPE 2021](https://www.iguazio.com/events/odsc-europe/): ODSC Europe Virtual Conference 2021 is one of the largest applied data science conferences and Iguazio will be there! Attend...
- [Expert Panel: AI for Connected Vehicles](https://www.iguazio.com/events/expert-panel-ai-for-connected-vehicles/): In this panel, experts from NetApp, NVIDIA, Strategy Analytics and Iguazio will discuss the main use cases, and also examine...
- [Webinar: "Activate Data: Data Science Innovation with MongoDB & Iguazio"](https://www.iguazio.com/events/mlops-live-webinar-15-activate-data-data-science-innovation-with-mongodb-iguazio/): The MLOps Live Webinar Series is a complimentary webcast where you will learn how to manage and automate machine learning...
- [KubeCon NA](https://www.iguazio.com/events/kubecon-na/): The Cloud Native Computing Foundation’s flagship conference gathers adopters and technologists from leading open source and cloud native communities virtually.
- [Toronto Machine Learning Summit](https://www.iguazio.com/events/toronto-machine-learning-summit/): The Toronto Machine Learning Summit (TMLS) is a uniquely interactive experience with a community of over 9,000 active members that...
- [NetApp INSIGHT 2020](https://www.iguazio.com/events/netapp-insight-2020/): NetApp INSIGHT 2020 is a fully digital, totally immersive event which explores how NetApp can help customers unlock the best...
- [Data Science Salon | Applying AI & ML to Healthcare, Finance & Technology](https://www.iguazio.com/events/data-science-salon-media-advertising-entertainment/): The data science salon is a unique vertical focused conference which grew into a diverse community of senior data science,...
- [MLOps Series Library](https://www.iguazio.com/events/mlops-in-flip-flops/): Access all past MLOps Live webinars to learn all about MLOps from industry experts and thought leaders.
- [MLOps Live Webinar #15: 'Automated Model Management for CPG Trade Effectiveness with Tredence'](https://www.iguazio.com/events/mlops-live/): The MLOps Live Webinar Series is a complimentary webcast where you will learn how to manage and automate machine learning...
- [MLOps NYC](https://www.iguazio.com/events/mlops-nyc/): MLOps NYC 2019 gathered industry leaders from companies like Netflix, Google, Twitter and Uber to share their biggest MLOps challenges...
- [KubeCon Europe](https://www.iguazio.com/events/london-in-february/): The Cloud Native Computing Foundation’s virtual conference will gather adopters and technologists from leading open source and cloud-native communities.
- [NVIDIA GTC: Deep Learning & AI Conference](https://www.iguazio.com/events/nvidia-gtc/): NVIDIA's GPU Technology Conference (GTC) is the must attend digital event for developers, researchers, engineers, and innovators looking to enhance...

---


## PRs

- [McKinsey & Company Acquires Iguazio to Accelerate & Scale Enterprise AI](https://www.iguazio.com/news-events/pr/mckinsey-company-acquires-iguazio-to-accelerate-scale-enterprise-ai/): Acquisition will enable AI's full power and potential to be realized across commercial, social, and environmental initiatives.
- [Sense Selects Iguazio for AI Chatbot Automation with AWS, Snowflake and NVIDIA](https://www.iguazio.com/news-events/pr/sense-selects-iguazio-for-ai-chatbot-automationwith-aws-snowflake-and-nvidia/): Sense will use the Iguazio MLOps platform for a large range of AI products, beginning with the Sense AI Chatbot.
- [Iguazio Partners with Snowflake to Automate and Accelerate MLOps](https://www.iguazio.com/news-events/pr/iguazio-partners-with-snowflake-to-automate-and-accelerate-mlops/): Read more about Iguazio Partners with Snowflake to Automate and Accelerate MLOps
- [LATAM Airlines Chooses Iguazio to Operationalize Machine Learning](https://www.iguazio.com/news-events/pr/latam-airlines-group-selects-iguazio-to-operationalize-machine-learning-as-part-of-its-post-pandemic-innovation-strategy/): Read more about LATAM Airlines Chooses Iguazio to Operationalize Machine Learning
- [Iguazio Partners with Pure Storage to Operationalize AI Production-First](https://www.iguazio.com/news-events/pr/pure-storage-and-iguazio-form-strategic-partnership-to-operationalize-ai-for-enterprises-taking-a-production-first-approach/): Read more about Iguazio Partners with Pure Storage to Operationalize AI Production-First
- [Iguazio MLOps Platform Now Supports Amazon FSx for NetApp ONTAP](https://www.iguazio.com/news-events/pr/iguazio-mlops-platform-now-supports-amazon-fsx-for-netapp-ontap/): Read more about Iguazio MLOps Platform Now Supports Amazon FSx for NetApp ONTAP
- [Iguazio MLOps Platform Launches in AWS Marketplace](https://www.iguazio.com/news-events/pr/iguazio-mlops-platform-launches-in-aws-marketplace/): Read more about Iguazio MLOps Platform Launches in AWS Marketplace
- [Boston Limited and Iguazio Partner to Operationalize AI for the Enterprise](https://www.iguazio.com/news-events/pr/boston-limited-and-iguazio-partner-to-operationalize-ai-for-the-enterprise/): Enabling both companies to extend their offerings to enterprises across industries looking to bring data science into real life applications.
- [Iguazio Announces First-Ever ‘MLOps for Good’ Virtual Hackathon](https://www.iguazio.com/news-events/pr/iguazio-announces-first-ever-mlops-for-good-virtual-hackathon/): With the mission to foster projects that can immediately impact real-world issues, Iguazio, partners Microsoft, MongoDB, and sponsor Aztek seek.
- [Iguazio Launches Integrated Feature Store to Accelerate AI Deployment](https://www.iguazio.com/news-events/pr/iguazio-launches-the-first-integrated-feature-store-within-its-data-science-platform-to-accelerate-deployment-of-ai-in-any-cloud-environment/): The first production-ready integrated solution for enterprises to catalog, store and share features centrally, and use them to manage AI applications.
- [Sheba Medical Center Partners with Iguazio for Real-Time COVID-19 AI](https://www.iguazio.com/news-events/pr/sheba-medical-center-inks-strategic-agreement-with-iguazio-to-deliver-real-time-ai-for-covid-19-patient-treatment-optimization/): Read more about Sheba Medical Center Inks Strategic Agreement with Iguazio to Deliver Real-Time AI for COVID-19.
- [Iguazio Achieves AWS Outposts Ready Status to Accelerate AI in Hybrid](https://www.iguazio.com/news-events/pr/iguazio-achieves-aws-outposts-ready-designation-to-help-enterprises-accelerate-ai-deployment-in-hybrid-environments/): The new seamless integration of Iguazio’s technology with AWS outposts allows customers to build ML pipelines in weeks instead of months.
- [PadSquad Deploys the Iguazio Data Science Platform to Predict Ad Performance in Real-Time](https://www.iguazio.com/news-events/pr/padsquad-deploys-the-iguazio-data-science-platform-to-predict-ad-performance-in-real-time/): Discover how PadSquad utilizes the Iguazio Data Science Platform to predict ad performance in real-time, enhancing marketing strategies.
- [Faktion and Iguazio Bring Data Science to Production for Smart Mobility](https://www.iguazio.com/news-events/pr/faktion-iguazio-bring-data-science-to-production-smart-mobility-customers/): The partnership enables both companies to provide AI infrastructure and services to smart mobility companies looking to harness big data.
- [SFL Scientific and Iguazio Partner to Accelerate Custom AI Development](https://www.iguazio.com/news-events/pr/sfl-scientific-and-iguazio-partner-to-speed-up-custom-ai-development-for-fortune-1000-companies/): Read more about SFL Scientific and Iguazio Partner to Accelerate Custom AI Development
- [NetApp Deploys Iguazio to Run AI-Driven Digital Advisor on Active IQ](https://www.iguazio.com/news-events/pr/netapp-deploys-iguazio-to-run-ai-driven-digital-advisor-on-active-iq-2/): NetApp deploys the Iguazio platform to boost the infrastructure behind its Active IQ solution, responding in real-time to 10 trillion data points.
- [Iguazio Becomes Certified for NVIDIA DGX-Ready Software Program](https://www.iguazio.com/news-events/pr/iguazio-becomes-certified-for-nvidia-dgx-ready-software-program/): Iguazio achieves NVIDIA DGX-Ready certification, validating its software for high-performance AI applications.
- [Iguazio and NetApp Collaborate to Accelerate Deployment of AI Applications](https://www.iguazio.com/news-events/pr/iguazio-and-netapp-collaborate-to-accelerate-deployment-of-ai-applications/): Partnership that provides enterprises with a simple, end-to-end solution for developing, deploying and managing AI applications.
- [Iguazio Deployed by Payoneer to Prevent Fraud with Real-time Machine Learning](https://www.iguazio.com/news-events/pr/iguazio-deployed-by-payoneer-to-prevent-fraud-with-real-time-machine-learning/): Payoneer uses Iguazio to move from detection to prevention of fraud prevention with predictive machine learning models served in real-time.
- [Iguazio Raises $24M to Accelerate Growth of Its Data Science Platform](https://www.iguazio.com/news-events/pr/iguazio-raises-24m-to-accelerate-growth-and-global-penetration-of-its-data-science-platform/): Iguazio Raises $24M to Accelerate Growth and Global Penetration of its Data Science Platform enabling a wide range of industries.
- [PICSIX Launches Investigative Intelligence Platform Powered by Iguazio](https://www.iguazio.com/news-events/pr/picsix-iguazio/): PICSIX uses Iguazio to provide an AI-based platform addressing the ever-changing threats to homeland security and public safety
- [Iguazio Expands Serverless To Scale-out Machine Learning and Analytics Workloads](https://www.iguazio.com/news-events/pr/iguazio-expands-serverless/): New serverless capabilities in Iguazio’s Platform enable on-demand resource consumption, elastic scaling, and simpler ML pipelines
- [MLOps NYC19 Conference to Promote the Standardization of Machine Learning Operations](https://www.iguazio.com/news-events/pr/mlops-nyc19-conference-to-promote-the-standardization-of-machine-learning-operations/): MLOps NYC19 will reflect the current state of machine learning operations with accomplished industry leaders sharing insights and experiences.
- [Iguazio to Operationalize Data Science and AI on Azure and Azure Stack](https://www.iguazio.com/news-events/pr/iguazio-to-operationalize-data-science-and-ai-on-azure-and-azure-stack/): Enabling Leading Enterprises to Bring Machine Learning into Business Applications and Remove AI Project Complexity
- [Iguazio’s Platform Scales NVIDIA GPU-Accelerated Deployments](https://www.iguazio.com/news-events/pr/iguazio-platform-scales-nvidia-gpu-accelerated-deployments/): Samsung SDS Integrates Serverless and Big Data to Automate Machine Learning Application Scaling and Productization
- [Samsung SDS Invests in Iguazio to Boost Cloud Services](https://www.iguazio.com/news-events/pr/samsung-sds-invests-in-iguazio-to-boost-cloud-services/): Samsung Adopts Iguazio’s Nuclio Serverless PaaS for Real-time Intelligent Applications
- [Iguazio Powers the Intelligent Edge for Smart Retail and IoT Solutions with Google Cloud](https://www.iguazio.com/news-events/pr/google-iguazio-intelligent-edge-retail/): Collaborating with Trax on a Kubernetes-powered hybrid cloud for real-time supply chain and intelligent operations
- [Iguazio's Nuclio Update Enables Serverless Agility for Real-Time Apps](https://www.iguazio.com/news-events/pr/iguazio-new-nuclio-release-enables-serverless-agility-enterprise-deploying-real-time-intelligent-applications/): Leading open source serverless framework now includes capabilities that enable faster end-to-end enterprise and IoT deployments.
- [Iguazio Hosts Serverless NYC: Enterprise Deployments and Case Studies](https://www.iguazio.com/news-events/pr/iguazio-hosts-serverless-nyc/): Iguazio Hosts Serverless NYC to Go Beyond the Hype, Presenting Enterprise Deployments and Real-Life Case Studies - find more here.
- [Equinix and Iguazio Collaborate to Drive Smart Mobility Vision](https://www.iguazio.com/news-events/pr/equinix-iguazio-collaborate-drive-smart-mobility-vision/): Hybrid cloud solution enables leading Asian ride-hailing applications to deliver real time and event-driven insights
- [Iguazio's Real-Time Serverless Framework Now Available for Enterprises](https://www.iguazio.com/news-events/pr/real-time-serverless-available-for-enterprise/): Nuclio enterprise edition gains momentum with leading cloud providers, as well as in on prem deployments and at the edge
- [PickMe Deploys Iguazio’s Platform for Real-Time Heatmaps, Fraud Detection](https://www.iguazio.com/news-events/pr/pickme-selects-iguazio/): Sri Lanka’s highest performing mobility app uses iguazio to maintain its competitive edge and operational efficiency
- [iguazio Featured in CRN’s 2018 Partner Program Guide](https://www.iguazio.com/news-events/pr/crn-partner-program-guide/): Annual Guide Recognizes the IT Channel’s Top Partner Programs
- [iguazio Extends Global Reach with New Channel Partner Program](https://www.iguazio.com/news-events/pr/iguazio-extends-global-reach-with-new-channel-partner-program/): Continuous Data Platform Now Available to More Enterprises Through Systems Integrators, VARs, OEMs
- [Unified Data Platform Provider iguazio Opens APAC Headquarters in Singapore](https://www.iguazio.com/news-events/pr/unified-data-platform-provider-apac-hq/): Demand for iguazio’s Hybrid Cloud and Edge Solutions Drives Continued Global Expansion opening of its APAC regional headquarters in Singapore
- [iguazio Debuts the nuclio Serverless Platform for Multi-Cloud and Edge Deployments](https://www.iguazio.com/news-events/pr/iguazio-debuts-serverless-platform/): nuclio expands iguazio’s data platform to provide a complete cloud experience that allows faster, flexible deployment in the cloud.
- [iguazio Announces General Availability of Its Unified Data Platform](https://www.iguazio.com/news-events/pr/iguazio-announces-general-availability-unified-data-platform/): Early customer adoption includes Grab, the Largest Ride-Hailing Service in Southeast Asia
- [Grab, Southeast Asia’s #1 Ride-Hailing Service, Selects iguazio’s Unified Data Platform](https://www.iguazio.com/news-events/pr/grab-south-east-asias-1-ride-hailing-service-selects-iguazios-unified-data-platform/): Largest Ride-Hailing Service in Southeast Asia uses Iguazio to Ingest, Enrich and Analyze Data for Continuous Analytics and more.
- [Iguazio Raises $33M in Series B as Leader in Real-Time Analytics, Edge Data](https://www.iguazio.com/news-events/pr/iguazio-leader-real-time-analytics-edge-data-platforms-raises-33m-series-b-funds/): New round includes strategic investors from financial services, IoT and service providers following successful early deployments
- [iguazio Selected as a Gartner  Cool Vendor in Data Management, 2017](https://www.iguazio.com/news-events/pr/iguazio-selected-as-a-gartner-cool-vendor-in-data-management-2017/): Delivering continuous analytics and offering greater simplicity, performance, security and agility for next generation applications
- [iguazio Demos Industry’s First Integrated Real-time Continuous Analytics Solution](https://www.iguazio.com/news-events/pr/iguazio-demos-industrys-first-integrated-real-time-continuous-analytics-solution/): Complete re-thinking of the traditional data pipeline reduces time-to-insights from hours to seconds ream-time Continuous Analytics Solution
- [iguazio Announces the World’s Fastest, Simplest and Lowest-Cost Enterprise Data Cloud](https://www.iguazio.com/news-events/pr/iguazio-announces-the-worlds-fastest-simplest-and-lowest-cost-enterprise-data-cloud/): Delivers 100x faster performance, Simplest and 10x lower cost for on-premises and hybrid cloud deployments
- [Iguazio Unveils World’s First Virtualized Data Services Architecture](https://www.iguazio.com/news-events/pr/iguazio-unveils-worlds-first-virtualized-data-services-architecture/): Reveals Details of Extremely Efficient Architecture that Seamlessly Accelerates Spark and Hadoop, Busts Silos and Ends ETL
- [Iguaz.Io raises $15 million in series a funding to disrupt big data storage](https://www.iguazio.com/news-events/pr/iguaz-io-raises-15-million-in-series-a-funding-to-disrupt-big-data-storage/): Iguaz.io, a provider of innovative data management and storage solutions for Big Data, IoT and cloud applications announced funding round.

---


## Sessions

- [Agentic AI Frameworks: Bridging Foundation Models and Business Impact](https://www.iguazio.com/sessions/agentic-ai-frameworks-bridging-foundation-models-and-business-impact/): Explore how Agentic AI frameworks can enhance decision-making and operational efficiency across industries.
- [Deploying Gen AI in Production with NVIDIA NIM & MLRun](https://www.iguazio.com/sessions/deploying-gen-ai-in-production-with-nvidia-nim-mlrun/): Learn to deploy Gen AI applications efficiently using NVIDIA NIM and MLRun for streamlined AI orchestration.
- [Gen AI for Marketing - From Hype to Implementation](https://www.iguazio.com/sessions/gen-ai-for-marketing-from-hype-to-implementation/): Transform your marketing strategy with Gen AI insights and actionable steps for effective implementation and risk management.
- [Building Scalable Customer-Facing Gen AI Applications Effectively & Responsibly](https://www.iguazio.com/sessions/building-scalable-customer-facing-gen-ai-applications-effectively-responsibly/): Learn how to Building Scalable Customer-Facing Gen AI Applications Effectively & Responsibly in our session.
- [Implementing Gen AI in Highly Regulated Environments](https://www.iguazio.com/sessions/implementing-gen-ai-in-highly-regulated-environments/): Discover unique challenges and solutions for deploying Gen AI in regulated industries to ensure compliance and performance.
- [Transforming Enterprise Operations with Gen AI](https://www.iguazio.com/sessions/transforming-enterprise-operations-with-gen-ai/): Explore how Gen AI enhances operations, focusing on real-world applications across manufacturing and supply chains.
- [Improving LLM Accuracy & Performance](https://www.iguazio.com/sessions/improving-llm-accuracy-performance/): Uncover best practices for improving LLM accuracy and performance while managing risk and costs effectively
- [LLM Validation & Evaluation](https://www.iguazio.com/sessions/llm-validation-evaluation/): Learn from our experts how to validate and evaluate LLMs effectively with automation strategies for improved accuracy.
- [Implementing a Gen AI Smart Call Center Analysis App](https://www.iguazio.com/sessions/implementing-a-gen-ai-smart-call-center-analysis-app/): Learn how to build a Gen AI-driven smart call center analysis app tailored for effective customer engagement.
- [GenAI for Financial Services](https://www.iguazio.com/sessions/genai-for-financial-services/): Generative AI has sparked the imagination with the explosion of tools, highlighting the importance of LLMs as the basis for modern AI.
- [Sheba Medical Center Improves Patient Outcomes and Experiences with AI](https://www.iguazio.com/sessions/sheba-medical-center-healthcare-ai/): Sheba Medical Center is driving a digital transformation by unifying real-time and historic data from different sources and more.
- [How to Build an Automated AI ChatBot](https://www.iguazio.com/sessions/how-to-build-an-automated-ai-chatbot/): Gennaro describe how he and his team have built and perfected this chatbot, what their ML pipeline looks like behind the scenes.
- [Demo: LLM Call Center Analysis with MLRun](https://www.iguazio.com/sessions/demo-llm-call-center-analysis-with-mlrun/): We showcase how to use LLMs to turn audio files of conversations between customers and agents at a call center into valuable data.
- [MLOps for Gen AI in the Enterprise](https://www.iguazio.com/sessions/mlops-for-gen-ai-the-live-webinar-series/): We will explore the effective integration of GenAI into real-time business applications, along with associated challenges & solutions.
- [MLOps for Generative AI](https://www.iguazio.com/sessions/mlops-for-generative-ai/): The influx of new tools spark the imagination and highlight the importance of Generative AI and foundation models as the basis for modern AI applications.
- [MLOps for LLMs](https://www.iguazio.com/sessions/mlops-for-llms/): Unlocking Generative AI's potential. Learn MLOps strategies for deploying and optimizing Hugging Face models in real business environments efficiently.
- [How Seagate Runs Advanced Manufacturing at Scale](https://www.iguazio.com/sessions/how-seagate-runs-advanced-manufacturing-at-scale/): Like most enterprises, Seagate was facing numerous challenges around AI and MLOps, at their hard disk advanced manufacturing sites.
- [HCI’s Journey to MLOps Efficiency](https://www.iguazio.com/sessions/hcis-journey-to-mlops-efficiency/): This is just what the data science team at HCI is doing. In this session, Jiri will be sharing enterprise secrets to establishing efficient systems for ML/AI.
- [How to Easily Deploy Your Hugging Face Model to Production at Scale](https://www.iguazio.com/sessions/how-to-easily-deploy-your-hugging-face-model-to-production-at-scale/): Seems like almost everyone uses Hugging Face to simplify and reuse advanced models and work collectively as a community.
- [Breaking AI Bottlenecks with Iguazio + Amazon FSx for NetApp ONTAP](https://www.iguazio.com/sessions/fsx/): Learn how to overcome AI bottlenecks using Iguazio and Amazon FSx for faster data processing and enhanced analytics.
- [From AutoML to AutoMLOps: Automated Logging & Tracking of ML](https://www.iguazio.com/sessions/automl-to-automlops-automated-logging-and-tracking-of-ml/): In this session, Yaron and Guy outline the challenges, describe open-source tools available for Auto-MLOps, and finish off with a live demo.
- [Best Practices for Succeeding with MLOps](https://www.iguazio.com/sessions/best-practices-for-succeeding-with-mlops/): Discover effective MLOps strategies and best practices to streamline your machine learning operations and enhance collaboration.
- [Predicting 1st Day Churn with Real-Time AI](https://www.iguazio.com/sessions/predicting-1st-day-churn-with-real-time-ai/): his On-Demand MLOps Live Webinar goes beyond theory, with industry leaders sharing challenges and practical solutions.
- [LATAM Customer Testimonial](https://www.iguazio.com/sessions/latam-customer-testimonial/): Learn how Latam Airlines Group uses Iguazio across to improve pilot training and fraud detection, creating a safer and faster way to operate their models.
- [Git Based CI/CD for ML](https://www.iguazio.com/sessions/git-based-cicd-for-ml/): In this session, Yaron Haviv, discusses how to enable continuous delivery of machine learning to production using Git-based ML pipelines.
- [Scaling NLP Pipelines at S&P Global (IHS Markit)](https://www.iguazio.com/sessions/scaling-nlp-pipelines-at-ihs-markit/): The data science team at S&P Global (IHS Markit) share practical advice on building sophisticated NLP pipelines that work at scale.
- [Building a Real-Time ML Pipeline with a Feature Store](https://www.iguazio.com/sessions/building-a-real-time-ml-pipeline-with-a-feature-store/): Join us to explore real-time ML pipelines using a feature store, enhancing your machine learning workflows with efficiency.
- [Automated Model Management for CPG Trade Effectiveness](https://www.iguazio.com/sessions/automated-model-management-for-cpg-trade-effectiveness/): Leading consumer packaged goods companies excel at executing large-scale trade promotions, but struggle in running precisely the right promotion at scale.
- [Automating & Governing AI Over Production Data on Azure](https://www.iguazio.com/sessions/automating-governing-ai-over-production-data-on-azure/): Many enterprises today face numerous challenges around handling data for AI/ML. They find themselves having to manually extract datasets.
- [How Feature Stores Accelerate & Simplify Deployment of AI to Production](https://www.iguazio.com/sessions/how-feature-stores-accelerate-simplify-deployment-of-ai-to-production/): In this live session we will discuss how feature stores can be used to accelerate AI deployment, show a demo and present a customer use case.
- [Simplifying Deployment of ML in Federated Cloud and Edge Environments](https://www.iguazio.com/sessions/simplifying-deployment-of-ml-in-federated-cloud-and-edge-environments/): Unlock AI adoption: Hybrid solutions for data challenges, simplified cloud-edge deployment. Learn practical insights in ML edge apps, cloud-edge harmony.
- [Handling Large Datasets in Data Preparation & ML Training Using MLOps](https://www.iguazio.com/sessions/handling-large-datasets-in-data-preparation-ml-training-using-mlops/): Learn how to use Dask, Kubernetes, and MLRun to scale your data prep and ML training with ease.
- [Quadient Customer Testimonial](https://www.iguazio.com/sessions/quadient-customer-testimonial/): Experience Quadient's customer journey firsthand with a testimonial at Iguazio. Find inspiration for your business transformation.
- [NetApp Customer Testimonial](https://www.iguazio.com/sessions/netapp-customer-testimonial/): Discover the powerful testimonial of NetApp at Iguazio! Unlock transformative insights and strategies today.
- [NVIDIA on Industrializing Enterprise AI with the Right Platform](https://www.iguazio.com/sessions/nvidia-on-industrializing-enterprise-ai-with-the-right-platform/): Enterprises need a platform that brings together tools to streamline data science workflow and bring innovative concepts into production sooner.
- [Siemens on the Importance of Data Storytelling in Shaping a Data Science Product](https://www.iguazio.com/sessions/siemens-on-the-importance-of-data-storytelling-in-shaping-a-data-science-product/): A deep dive into storytelling with data and how to make sure all your hard work on developing the right model.
- [NetApp's Michael Oglesby on Building ML Pipelines Over Federated Data](https://www.iguazio.com/sessions/building-ml-pipelines-over-federated-data-compute-environments/): Learn more about constructing ML pipelines across federated data and computing environments in our session.
- [Product Madness (an Aristocrat co.) on Predicting 1st-Day Churn in Real Time](https://www.iguazio.com/sessions/product-madness-an-aristocrat-co-on-predicting-1st-day-churn-in-real-time/): Hear from Product Madness about how technology and new work processes can help the gaming and mobile app industries.
- [How to Deal With Concept Drift in Production with MLOps Automation](https://www.iguazio.com/sessions/how-to-deal-with-concept-drift-in-production-with-mlops-automation/): How to detect and handle problems that arise when models lose their accuracy and how to implement concept drift detection in production.
- [S&P Global’s Ganesh Nagarathnam on Bringing ML Pipelines to Production](https://www.iguazio.com/sessions/sp-globals-ganesh-nagarathnam-on-bringing-ml-pipelines-to-production/): Gain insights from S&P Global on operationalizing ML pipelines and aligning business needs with data-driven solutions.
- [Greg Hayes on Uniting Data Scientists, Engineers, and DevOps with MLOps](https://www.iguazio.com/sessions/breaking-the-silos-between-data-scientists-engineers-and-devops-with-new-mlops-practices/): Learn with Greg Hayes how Ecolab is accelerating the deployment of AI applications by using new MLOps methodologies.
- [NetApp’s Shankar Pasupathy on Building Scalable Predictive Maintenance](https://www.iguazio.com/sessions/how-to-build-a-predictive-maintenance-and-actionable-intelligence-solution-at-scale/): How built a solution for predictive maintenance and actionable intelligence that responds in real time to 10 trillion data points per month.
- [Microsoft & GitHub on Git-Based CI / CD for Machine Learning & MLOps](https://www.iguazio.com/sessions/git-based-ci-cd-for-machine-learning-mlops/): Learn about using Git to enable continuous delivery of ML to production, enable controlled collaboration across ML teams, and solve rigorous MLOps needs.
- [Quadient’s Jason Evans on Saving Time & Costs Bringing AI to Production](https://www.iguazio.com/sessions/quadients-jason-evans-on-saving-costs-bringing-ai-to-production/): Learn how industry leaders save costs and get to market faster by leveraging ML pipeline automation and open-source technology.

---


## MLOps Terminologies

- [AI Proof of Concept](https://www.iguazio.com/glossary/ai-proof-of-concept/): A small-scale, low-risk project designed to test whether an AI solution can deliver business value in a particular context.
- [Knowledge Distillation](https://www.iguazio.com/glossary/knowledge-distillation/): Knowledge distillation transfers knowledge from a large, complex model (called the teacher) to a smaller, simpler model (called the student).
- [Hyperpersonalization in AI](https://www.iguazio.com/glossary/hyperpersonalization-in-ai/): Hyperpersonalization in AI is when AI, ML and real-time data analytics are used to deliver highly individualized experiences, content, or services to users.
- [Self Reflection in LLMs](https://www.iguazio.com/glossary/self-reflection-in-llms/): Self-reflection in an LLM refers to a model's ability to assess, critique and improve its own outputs.
- [Mixture of Experts](https://www.iguazio.com/glossary/mixture-of-experts/): Mixture of Experts (MoE) in AI is a technique that uses multiple specialized models, called "experts," to solve complex problems efficiently.
- [Arithmetic Intensity](https://www.iguazio.com/glossary/arithmetic-intensity/): A frontier model is a highly advanced, large-scale AI model that pushes the boundaries of AI in areas like NLP, image generation, video and coding.
- [Frontier model](https://www.iguazio.com/glossary/frontier-model/): A frontier model is a highly advanced, large-scale AI model that pushes the boundaries of AI in areas like NLP, image generation, video and coding.
- [Context Window](https://www.iguazio.com/glossary/context-window/): The context window for an LLM refers to the maximum number of tokens the model can process in a single interaction.
- [Excessive Agency](https://www.iguazio.com/glossary/excessive-agency/): Excessive agency in LLMs refers to AI making decisions beyond intent, posing risks like bias, misinformation, and loss of control.
- [Reasoning Engine](https://www.iguazio.com/glossary/reasoning-engine/): Learn how to use an AI reasoning engine and why, overcoming hallucinations, components, use cases & how it works.
- [LLM Orchestration](https://www.iguazio.com/glossary/llm-orchestration/): Learn what LLM orchestration is, its key components, and the benefits it offers for optimizing large language model workflows and performance.
- [AI Scalability](https://www.iguazio.com/glossary/ai-scalability/): Discover what AI scalability means, explore key best practices, and tackle common challenges to optimize AI performance and growth.
- [LLM Tracing](https://www.iguazio.com/glossary/llm-tracing/): LLM tracing is the practice of tracking and understanding the decision-making and thought processes within LLMs as they generate responses.
- [Human in the Loop](https://www.iguazio.com/glossary/human-in-the-loop/): "Human in the loop" (HITL) is the process that blends human intervention in an automated or semi-automated AI/ML system.
- [LLM Monitoring](https://www.iguazio.com/glossary/what-is-llm-monitoring/): LLM Monitoring is the set of practices and tools used to track, validate and maintain the performance, safety and quality of LLMs. Learn how.
- [Random Forest](https://www.iguazio.com/glossary/what-is-random-forest/): Understand random forest - An ensemble learning method that combines results from multiple decision trees during training.
- [Prompt Management](https://www.iguazio.com/glossary/what-is-prompt-management/): Understand prompt management - Optimizing and maintaining the prompts used to interact with LLMs for quality, consistency and scalability.
- [LLM Customization](https://www.iguazio.com/glossary/llm-customization/): Explore LLM customization, its importance & applications in fine-tuning, prompt engineering, RAG, RAFT, agents, and performance optimization
- [LLM as a Judge](https://www.iguazio.com/glossary/llm-as-a-judge/): "LLM as a judge" refers to the use of LLMs to evaluate content, responses, or performances, including the performance of other AI models.
- [Chain-of-Thought Prompting](https://www.iguazio.com/glossary/chain-of-thought-prompting/): Chain-of-Thought (CoT) Prompting encourages the model to break down the problem into a series of smaller, logical steps
- [LLM Embeddings](https://www.iguazio.com/glossary/llm-embeddings/): LLM embeddings are vector representations of words, phrases, or entire texts generated by language models. Discover how they work.
- [Gen AI App](https://www.iguazio.com/glossary/gen-ai-app/): Gen AI apps use generative AI technologies and leverage advanced ML models like LLMs, GANs and VAEs, to create or generate new content.
- [AI Infrastructure](https://www.iguazio.com/glossary/ai-infrastructure/): LLM temperature is a parameter that influences the language model’s output, determining whether the output is more creative or predictable.
- [Diffusion Models](https://www.iguazio.com/glossary/diffusion-models/): Diffusion models are powerful generative models that are able to produce high-quality, diverse samples of image, audio or text.
- [Generative Agents](https://www.iguazio.com/glossary/generative-agents/): Generative agents are software entities that use generative models to simulate and mimic human behavior and responses.
- [LLM Optimization](https://www.iguazio.com/glossary/llm-optimization/): LLM temperature is a parameter that influences the language model’s output, determining whether the output is more creative or predictable.
- [LLM Temperature](https://www.iguazio.com/glossary/llm-temperature/): LLM temperature is a parameter that influences the language model’s output, determining whether the output is more creative or predictable.
- [LLM Agents](https://www.iguazio.com/glossary/llm-agents/): An on-premise AI platform runs AI services and applications within the organization's physical environment, rather than on the cloud.
- [On-Premise AI Platform](https://www.iguazio.com/glossary/on-premise-ai-platform/): An on-premise AI platform runs AI services and applications within the organization's physical environment, rather than on the cloud.
- [False Positive Rate](https://www.iguazio.com/glossary/false-positive-rate/): Get insights into the false positive rate in machine learning, its implications, and how to assess model performance.
- [True Positive Rate](https://www.iguazio.com/glossary/true-positive-rate/): True positive rate (TPR) is a performance metric used to evaluate the effectiveness of binary classification models in machine learning
- [RLHF](https://www.iguazio.com/glossary/rlhf/): RLHF is an AI/ML model training approach that combines reward-based reinforcement learning methods and human-generated feedback.
- [Fine-Tuning LLMs](https://www.iguazio.com/glossary/fine-tuning/): Fine-tuning LLMs (Large Language Models) is the process of adapting a pre-trained language model to a specific task or dataset.
- [Prompt Engineering](https://www.iguazio.com/glossary/prompt-engineering/): Prompt engineering is the process of formulating inputs (prompts) to an AI model (usually an LLM) to achieve the desired outputs.
- [LLM Hallucinations](https://www.iguazio.com/glossary/llm-hallucination/): LLMs can sometimes produce outputs that are coherent and grammatically correct but factually incorrect or nonsensical.
- [Auto-Regressive Models](https://www.iguazio.com/glossary/auto-regressive-models/): Auto-regressive models are models used in time series datasets that predict future values based on past values.
- [AI Tokenization](https://www.iguazio.com/glossary/ai-tokenization/): Dive into AI tokenization, explore its usefulness in the AI ecosystem, and understand the advantages it brings to AI applications.
- [Model Behavior](https://www.iguazio.com/glossary/model-behavior/): Model behavior is the way in which a trained ML model makes predictions and decisions when it is exposed to new data.
- [Baseline Models](https://www.iguazio.com/glossary/baseline-models/): Baseline models are simple models that serve as the basis for evaluating the performance of more complex models
- [ML Stack](https://www.iguazio.com/glossary/ml-stack/): An ML Stack is the entire collection of technologies and frameworks used throughout ML development, deployment and management.
- [LLMOps](https://www.iguazio.com/glossary/llmops/): LLMOps refers to the set of practices and tools used to manage, streamline, and operationalize large language models, is a portmanteau of ‘LLM’ and ‘MLOps’.
- [Transfer Learning](https://www.iguazio.com/glossary/transfer-learning/): Transfer learning works by leveraging a pre-trained model's learned features and fine-tuning it on a new task-specific dataset.
- [Foundation Models](https://www.iguazio.com/glossary/foundation-models/): These models, often referred to as "base models" or "pre-trained models," have quickly become the building blocks of many advanced AI systems.
- [Large Language Models](https://www.iguazio.com/glossary/large-language-model-llms/): A large language model is an advanced AI system that has been trained on extensive amounts of data to understand and generate human-like language.
- [Data Ingestion](https://www.iguazio.com/glossary/data-ingestion/): Data ingestion for machine learning refers to the process of collecting and preparing data for use in machine learning models.
- [Data Pipeline Automation](https://www.iguazio.com/glossary/data-pipeline-automation/): Data pipeline automation is the process of automating the flow of data from one system or application to another.
- [Risk Management](https://www.iguazio.com/glossary/risk-management/): This article introduces the concept of risk management for ML and what technical risks come with ML models.
- [MLOps Governance](https://www.iguazio.com/glossary/mlops-governance/): Model governance is a supplementary component to MLOps that supports AI compliance and traceability by improving visibility and control over ML deployments
- [Model Evaluation](https://www.iguazio.com/glossary/model-evaluation/): This article defines model evaluation, discusses its importance, and introduces best practices on how to report on and perform the evaluation.
- [Holdout Dataset](https://www.iguazio.com/glossary/holdout-dataset/): When labels are present (i.e., in supervised learning,) the most simple and commonly used approach for model evaluation is a holdout dataset.
- [Overfitting](https://www.iguazio.com/glossary/overfitting/): Model overfitting is a statistical error in supervised ML, whereby the trained model fits the noise in the training data rather than its actual pattern.
- [Open Source Model](https://www.iguazio.com/glossary/open-source-model/): Open source has always been an integral part of AI. Open source refers to software for which the original source code is made publicly available.
- [Cross-Validation](https://www.iguazio.com/glossary/cross-validation/): An introduction to cross-validation, an overview of its benefits, and a walk through of when and how to use this technique in machine learning.
- [Automated Machine Learning](https://www.iguazio.com/glossary/automated-machine-learning/): Automated machine learning aims to simplify the entry requirements and ongoing resource requirements for developing real life AI solutions.
- [Recall](https://www.iguazio.com/glossary/recall/): Recall in ML: an introduction to this machine learning metric, a discussion of when to use it, and a walk-through of how to improve it.
- [Classification Threshold](https://www.iguazio.com/glossary/classification-threshold/): Classification is the set of algorithms that, together with regression, comprises supervised machine learning (ML).
- [Regression](https://www.iguazio.com/glossary/regression/): This article presents an introduction to ML regression, a review of the most common associated evaluation metrics, a walk-through of when to use regression
- [Model Training](https://www.iguazio.com/glossary/model-training/): This article presents an introduction to model training, a discussion of its importance, and a walk-through of how to train ML models during experimentation.
- [Continuous Validation](https://www.iguazio.com/glossary/continuous-validation/): A comprehensive definition of continuous validation in ML, a review of its importance, and a walk-through of the most common tools used in the field.
- [Model Tuning](https://www.iguazio.com/glossary/model-tuning/): What hyper parameters and model tuning are, why model tuning is important, and how to successfully tune your machine learning models.
- [Noise in ML](https://www.iguazio.com/glossary/noise-in-ml/): Understand the concept of noise in machine learning and its impact on model accuracy and decision-making, and how to handle.
- [Model Accuracy in Machine Learning](https://www.iguazio.com/glossary/model-accuracy-in-ml/): AI accuracy is the percentage of correct classifications that a trained machine learning model achieves.
- [Explainable AI](https://www.iguazio.com/glossary/explainable-ai/): Explainable AI (XAI) is a set of tools and methods that attempt to help humans understand the outputs of machine learning models.
- [Drift Monitoring](https://www.iguazio.com/glossary/drift-monitoring/): A part of the MLOps process, drift monitoring ensures model performance and relevance.
- [Image Processing Framework](https://www.iguazio.com/glossary/image-processing-framework/): Image processing is the series of operations aimed at improving the quality of images for computer vision tasks so they can be more predictive. 
- [Model Serving Pipeline](https://www.iguazio.com/glossary/model-serving-pipeline/): A machine learning (ML) model pipeline or system is a technical infrastructure used to automatically manage ML processes. Learn more here.
- [Kubernetes for MLOps](https://www.iguazio.com/glossary/kubernetes-for-mlops/): Kubernetes is the preferred MLOps tool to manage automated machine learning pipelines in a reproducible, safe, and scalable way.
- [GPU for Machine Learning](https://www.iguazio.com/glossary/gpu-for-machine-learning/): What a GPU is, how it works, how it compares with other typical hardware for ML, and how to select the best GPU for your application.
- [Machine Learning Infrastructure](https://www.iguazio.com/glossary/machine-learning-infrastructure/): A complete machine learning infrastructure famously involves a much wider variety of engineering and procedural components.
- [Model Retraining](https://www.iguazio.com/glossary/model-retraining/): ML model retraining is the MLOps capability to automatically and continuously retrain a machine learning model on a schedule or a trigger.
- [Model Deployment](https://www.iguazio.com/glossary/model-deployment/): Model deployment is the process of putting ML model into production, and it can be resource-intensive and time consuming.
- [Deep Learning Pipelines](https://www.iguazio.com/glossary/deep-learning-pipelines/): Bringing deep learning use cases to production is a particularly complex AI project, given the scale of data required.
- [Model Management](https://www.iguazio.com/glossary/model-management/): Model management is the component of MLOps that ensures a machine learning model is set up correctly.
- [Feature Vector](https://www.iguazio.com/glossary/feature-vector/): ML Glossary: A feature vector is an ordered list of numerical properties of observed phenomena.
- [Model Serving](https://www.iguazio.com/glossary/model-serving/): What production-grade model serving actually is, plus model serving use cases, tools, and model serving with Iguazio.
- [CI/CD for Machine Learning](https://www.iguazio.com/glossary/ci-cd-for-machine-learning/): In this article, we review the basics of a CI/CD pipeline and explain what implementing a CI/CD practice for ML entails
- [Feature Engineering](https://www.iguazio.com/glossary/feature-engineering/): Feature engineering selects and transforms the most relevant variables from raw data to create input features to machine learning models for inferencing.
- [Real Time ML](https://www.iguazio.com/glossary/real-time-ml/): Real-time machine learning applications need a real-time data pipeline to make time-critical decisions on fresh data.
- [Unsupervised Machine Learning](https://www.iguazio.com/glossary/unsupervised-ml/): Unsupervised machine learning algorithms can discover underlying features of a data set for further downstream processing and prediction tasks.
- [Kubeflow Pipelines](https://www.iguazio.com/glossary/kubeflow-pipelines/): Kubeflow pipelines is a platform for scheduling and orchestrating multi- and parallel-step ML workflows in a simple and robust way.
- [Machine Learning Lifecycle](https://www.iguazio.com/glossary/machine-learning-lifecycle/): The machine learning lifecycle is the cyclical process that data science projects follow, to bring business value from ML.
- [Model Monitoring](https://www.iguazio.com/glossary/model-monitoring/): Model performance monitoring is a basic operational task that is implemented after an AI model has been deployed.
- [Concept Drift](https://www.iguazio.com/glossary/concept-drift/): Concept drift is a natural part of an ML system. To ensure that models deliver value, ML teams need to build a drift-aware system.
- [Feature Store](https://www.iguazio.com/glossary/feature-store/): Feature stores are a central place to build, manage and share features across different teams in the organization.
- [Enterprise Data Science](https://www.iguazio.com/glossary/enterprise-data-science/): Enterprise data science combines data scientists, data engineers, IT teams, and more to generate value out of big data. 
- [Operationalizing ML](https://www.iguazio.com/glossary/operationalizing-machine-learning/): Operationalizing machine learning is one of the final stages before deploying and running an ML model in a production environment.
- [ML Pipeline](https://www.iguazio.com/glossary/machine-learning-pipeline/): A machine learning pipeline helps to streamline and speed up the process by automating these workflows and linking them together.
- [ML Pipeline Tools](https://www.iguazio.com/glossary/machine-learning-pipeline-tools/): A machine learning pipeline tool helps automate and streamline machine learning pipelines. Learn more on our page.

---


## Case Studies

- [HCI Builds a Mature Enterprise MLOps Practice to Deploy 73+ Financial Use Cases](https://www.iguazio.com/case-study/hci-mature-mlops-finance/): With Iguazio, Sense powers a wide range of AI products aimed at increasing the efficiency and scalability of their AI operations.
- [LATAM Airlines Group Builds an ML Factory to Generate Business Impact Across the Organization](https://www.iguazio.com/case-study/latam-airlines-group-builds-an-ml-factory-to-generate-business-impact-across-the-organization/): LATAM deploys over 40 AI services across commercial and operational departments to create business value and plan for the post-pandemic future.
- [Seagate Uses AI to Detect Defects on the Factory Floor, Reduce Cost and Improve Yield](https://www.iguazio.com/case-study/seagate-runs-advanced-manufacturing-at-scale/): Seagate transitioned from a manual inspection process via microscope to fully automated deep learning & computer vision inspection
- [Sense Scales Chatbot Automation and HR Tech](https://www.iguazio.com/case-study/sense-scales-chatbot-automation/): Sense leverages automation and AI to speed up the recruitment process, while delivering a hyper-personalized candidate experience.
- [Sense Personalizes the Recruitment Experience with AI Chatbot Automation](https://www.iguazio.com/case-study/sense-automates-and-personalizes-the-recruitment-experience-with-ai/): With Iguazio, Sense powers a wide range of AI products aimed at increasing the efficiency and scalability of their AI operations.
- [LATAM Airlines Group Drives Innovation with a Cross-Company AI Strategy](https://www.iguazio.com/case-study/latam-drives-innovation-with-a-cross-company-ai-strategy/): LATAM deploys over 40 AI services across commercial and operational departments to create business value and plan for the post-pandemic future.
- [LATAM Airlines Group Drives Innovation with a Cross Company AI Strategy](https://www.iguazio.com/case-study/latam-drives-innovation/): LATAM deploys over 40 AI services across commercial and operational departments to create business value and plan for the post-pandemic.
- [S&P Global Makes Engineering Documents Searchable and Indexable with NLP](https://www.iguazio.com/case-study/s-and-p-global-makes-engineering-documents-searchable-indexable-with-nlp/): S&P Global deploys semantic extraction on engineering documents to drive better decision making, processing thousands of PDF files and more.
- [Ecolab Reduces Time to AI Deployment from 12 months to 30 days](https://www.iguazio.com/case-study/ecolab-reduces-time-to-ai-deployment-from-12-months-to-30-days/): Hygiene technologies leader Ecolab brings data science to production with Microsoft Azure and Iguazio
- [Ecolab Deploys Predictive Risk Models](https://www.iguazio.com/case-study/ecolab/): Hygiene technologies leader Ecolab brings data science to production with Microsoft Azure and Iguazio
- [Sheba Medical Center Improves Patient Outcomes and Experiences with AI](https://www.iguazio.com/case-study/sheba-medical-center/): Sheba Medical Center is driving a digital transformation by rapidly deploying and monitoring AI models across clinical and logistical use cases.
- [NetApp Deploys Real-Time Predictive Maintenance](https://www.iguazio.com/case-study/netapp-deploys-real-time-predictive-maintenance/): NetApp deployed Iguazio at the core of Active IQ, analyzing 10 trillion data points in real time from storage sensors worldwide .
- [NetApp Deploys Real-Time Predictive Maintenance & Advanced Analytics](https://www.iguazio.com/case-study/netapp-deploys-real-time-predictive-maintenance-advanced-analytics/): NetApp deployed Iguazio at the core of Active IQ, analyzing data points from storage sensors worldwide to generate actionable intelligence.
- [PadSquad Predicts Ad Performance in Real Time Based on Multivariate Data](https://www.iguazio.com/case-study/padsquad/): PadSquad aggregates and processes real-time ad data, to optimize performance in real-time and deliver timely insights to customers
- [Quadient Saves Time and Costs Getting AI to Production](https://www.iguazio.com/case-study/quadient-case-study/): Quadient unifies and combines every single data type they work with, to help its clients deliver more meaningful customer experiences to consumers.
- [Ecolab Breaks the Silos Between Data Scientists, Engineers and DevOps with New MLOps Practices](https://www.iguazio.com/case-study/video-title-lorem-ipsum-dolor-sit-amet-consectetur-adipiscing-elit-sed-do-2/): How Ecolab is accelerating the deployment of AI applications by using new MLOps methodologies, leveraging microservices and more.
- [Payoneer Uses Real-Time AI for Fraud Prevention](https://www.iguazio.com/case-study/payoneer-case-study/): With a scalable and reliable fraud prediction and prevention model, fraud attacks are almost impossible on Payoneer.

---


## Solutions


---


## Q&A

- [Why should you combine traditional ML with LLMs?](https://www.iguazio.com/questions/why-should-you-combine-traditional-ml-with-llms/): Need to know Why should you combine traditional ML with LLMs? - Check our answer on Iguazio Q&A section.
- [What are the milestones of developing a multimodal agent?](https://www.iguazio.com/questions/what-are-the-milestones-of-developing-a-multimodal-agent/): Need to know What are the milestones of developing a multimodal agent? - Check our answer on Iguazio Q&A section.
- [Managing and Optimizing Costs in Production-Ready Generative AI](https://www.iguazio.com/questions/what-are-some-ways-to-manage-and-optimize-costs-when-deploying-generative-ai-in-production/): Need to know about Managing Costs When Deploying Generative AI Solutions - Check our answer on Iguazio Q&A section.
- [What training is recommended to upskill (gen) AI talent in the organization?](https://www.iguazio.com/questions/what-training-is-recommended-to-upskill-gen-ai-talent-in-the-organization/): Need to know What training is recommended to upskill (gen) AI talent in the organization? - Check our answer on Iguazio Q&A section.
- [What are some examples of how gen AI is impacting operational and business results?](https://www.iguazio.com/questions/what-are-some-examples-of-how-gen-ai-is-impacting-operational-and-business-results/): Need to know What are some examples of how gen AI is impacting operational and business results? - Check our answer on Iguazio Q&A section.
- [How do gen AI and traditional AI complement each other?](https://www.iguazio.com/questions/how-do-gen-ai-and-traditional-ai-complement-each-other/): Need to know How do gen AI and traditional AI complement each other? - Check our answer on Iguazio Q&A section.
- [What are the recommended steps for evaluating gen AI outputs?](https://www.iguazio.com/questions/what-are-the-recommended-steps-for-evaluating-gen-ai-outputs/): Need to know What are the recommended steps for evaluating gen AI outputs? - Check our answer on Iguazio Q&A section.
- [What Guardrails Can Be Implemented in (Gen) AI Pipelines?](https://www.iguazio.com/questions/guardrails-gen-ai-pipelines/): Need to know What Guardrails Can Be Implemented in (Gen) AI Pipelines? - Check our answer on Iguazio Q&A section.
- [Why is It Important to Monitor LLMs?](https://www.iguazio.com/questions/why-is-it-important-to-monitor-llms/): Need to know Why is It Important to Monitor LLMs? - Check our answer on Iguazio Q&A section.
- [RAG vs. Fine-tuning: When to Use Each One?](https://www.iguazio.com/questions/rag-vs-fine-tuning-when-to-use-each-one/): Need to know RAG vs. Fine-tuning: When to Use Each One? - Check our answer on Iguazio Q&A section.
- [What are some tips and steps for improving LLM prediction accuracy?](https://www.iguazio.com/questions/what-are-some-tips-and-steps-for-improving-llm-prediction-accuracy/): Need to know What are some tips and steps for improving LLM prediction accuracy? - Check our answer on Iguazio Q&A section.
- [Which gen AI smart call center app use cases are other companies implementing?](https://www.iguazio.com/questions/which-gen-ai-smart-call-center-app-use-cases-are-other-companies-implementing/): Need to know Which gen AI smart call center app use cases are other companies implementing? - Check our answer on Iguazio Q&A section.
- [What is hyper-personalization in gen AI?](https://www.iguazio.com/questions/what-is-hyper-personalization-in-gen-ai/): Need to know What is hyper-personalization in gen AI? - Check our answer on Iguazio Q&A section.
- [How can organizations address risks in gen AI?](https://www.iguazio.com/questions/how-can-organizations-address-risks-in-gen-ai/): Need to know How can organizations address risks in gen AI? - Check our answer on Iguazio Q&A section.
- [Can LLMs be implemented in non-English languages?](https://www.iguazio.com/questions/can-llms-be-implemented-in-non-english-languages/): Need to know Can LLMs be implemented in non-English languages? - Check our answer on Iguazio Q&A section.
- [What level of PII reduction accuracy in the AI pipeline is acceptable?](https://www.iguazio.com/questions/what-level-of-pii-reduction-accuracy-in-the-ai-pipeline-is-acceptable/): Need to know What level of PII reduction accuracy in the AI pipeline is acceptable? - Check our answer on Iguazio Q&A section.
- [What is a LLM pricing strategy?](https://www.iguazio.com/questions/what-is-a-llm-pricing-strategy/): Need to know What is a LLM pricing strategy? - Check our answer on Iguazio Q&A section. All you need about LLMops.
- [What are the risks of open source LLMs?](https://www.iguazio.com/questions/what-are-the-risks-of-open-source-llms/): Need to know What are the risks of open source LLMs? - Check our answer on Iguazio Q&A section.
- [How can LLMs help determine user intent?](https://www.iguazio.com/questions/how-can-llms-help-determine-user-intent/): Need to know How can LLMs help determine user intent? - Check our answer on Iguazio Q&A section.
- [What’s the best way to get related context to feed the prompt using similarity search?](https://www.iguazio.com/questions/whats-the-best-way-to-get-related-context-to-feed-the-prompt-using-similarity-search/): Need to know What’s the best way to get related context to feed the prompt using similarity search? - Check our answer on Iguazio Q&A section.
- [What are the privacy and security implications of using open source components in AI?](https://www.iguazio.com/questions/what-are-the-privacy-and-security-implications-of-using-open-source-components-in-ai/): Need to know What are the privacy and security implications of using open source components in AI? - Check our answer on Iguazio Q&A section.
- [What's the best way to extract data from video with generative AI?](https://www.iguazio.com/questions/whats-the-best-way-to-extract-data-from-video-with-generative-ai/): Need to know What's the best way to extract data from video with generative AI? - Check our answer on Iguazio Q&A section.
- [What’s the best way to perform search queries in documents with generative AI?](https://www.iguazio.com/questions/whats-the-best-way-to-perform-search-queries-in-documents-with-generative-ai/): Need to know What’s the best way to perform search queries in documents with generative AI? - Check our answer on Iguazio Q&A section.
- [Why is it important for data scientists and DevOps teams to collaborate and communicate around GenAI and MLOps?](https://www.iguazio.com/questions/why-is-it-important-for-data-scientists-and-devops-teams-to-collaborate-and-communicate-around-genai-and-mlops/): Need to know about the Importance of Collaboration in GenAI and MLOps - Check our answer on Iguazio Q&A section.
- [Addressing Scalability and Performance Challenges in Generative AI](https://www.iguazio.com/questions/whats-the-best-way-to-address-scalability-and-performance-challenges-for-a-generative-ai-app/): Need to know best way to address scalability and performance challenges for a generative AI app - Check our answer on Iguazio Q&A section.
- [What are some best practices for establishing thresholds to trigger a new iteration of a generative AI model with new prompts in the ML lifecycle?](https://www.iguazio.com/questions/what-are-some-best-practices-for-establishing-thresholds-to-trigger-a-new-iteration-of-a-generative-ai-model-with-new-prompts-in-the-ml-lifecycle/): What are best practices for establishing thresholds to trigger a new iteration of a generative AI model with new prompts in the ML lifecycle?
- [How can organizations implement guardrails to ensure ethical use of AI?](https://www.iguazio.com/questions/how-can-organizations-implement-guardrails-to-ensure-ethical-use-of-ai/): Need to know How can organizations implement guardrails to ensure ethical use of AI? - Check our answer on Iguazio Q&A section.
- [Is there a privacy issue or data leaking risk with custom models that utilize proprietary or public data?](https://www.iguazio.com/questions/is-there-a-privacy-issue-or-data-leaking-risk-with-custom-models-that-utilize-proprietary-or-public-data/): Need to know about privacy issue or data leaking risk with custom models that use public data - Check our answer on Iguazio Q&A section.
- [What is the key difference between fine tuning and embedding a foundational model?](https://www.iguazio.com/questions/what-is-the-key-difference-between-fine-tuning-and-embedding-a-foundational-model/): Need to know What is the key difference between fine tuning and embedding a foundational model? - Check our answer on Iguazio Q&A section.
- [What are the steps in the MLOps workflow that are specific to LLMs?](https://www.iguazio.com/questions/what-are-the-steps-in-the-mlops-workflow-that-are-specific-to-llms/): Need to know What are the steps in the MLOps workflow that are specific to LLMs? - Check our answer on Iguazio Q&A section.
- [How can LLMs be customized with company data?](https://www.iguazio.com/questions/how-can-llms-be-customized-with-company-data/): Need to know How can LLMs be customized with company data? - Check our answer on Iguazio Q&A section.
- [How can costs be optimized when deploying LLMs?](https://www.iguazio.com/questions/how-can-costs-be-optimized-when-deploying-llms/): Need to know How can costs be optimized when deploying LLMs? - Check our answer on Iguazio Q&A section.
- [Can MLRun be used with Amazon SageMaker?](https://www.iguazio.com/questions/can-mlrun-be-used-with-amazon-sagemaker/): Need to know Can MLRun be used with Amazon SageMaker? - Check our answer on Iguazio Q&A section.
- [MLRun vs. Seldon: What's the difference?](https://www.iguazio.com/questions/mlrun-vs-seldon-whats-the-difference/): Need to know MLRun vs. Seldon: What's the difference? - Check our answer on Iguazio Q&A section.
- [Can MLRun Support Models Built in AWS Dev Accounts for Upstream?](https://www.iguazio.com/questions/can-mlrun-support-models-built-in-an-aws-dev-account-and-promote-them-to-upstream-environments/): Need to know if Can MLRun support models built in an AWS Dev Account - Check our answer on Iguazio Q&A section.
- [<strong>Does MLRun orchestrate on a Kubernetes operator or use a classic Helm chart?</strong>](https://www.iguazio.com/questions/mlrun-k8s-or-helm-chart/): Does MLRun orchestrate on a Kubernetes operator or use a classic Helm chart? Check our answer on Iguazio Q&A section
- [Can MLRun utilize GPUs when running ML jobs?](https://www.iguazio.com/questions/can-mlrun-utilize-gpus-when-running-ml-jobs/): Can MLRun use GPUs? - Check our answer on Iguazio Q&A section.
- [<strong>How do I use MLRun for batch sizing?</strong>](https://www.iguazio.com/questions/batch-sizing/): How to use MLRun for batch sizing - Check our answer on Iguazio Q&A section.
- [<strong>How do I use MLRun for real-time streaming configuration?</strong>](https://www.iguazio.com/questions/how-do-i-use-mlrun-for-real-time-streaming-configuration/): Need to know <strong>How do I use MLRun for real-time streaming configuration?</strong> - Check our answer on Iguazio Q&A section.
- [How do MLRun and Iguazio plug into the ML ecosystem?](https://www.iguazio.com/questions/how-do-mlrun-and-iguazio-plug-into-the-ml-ecosystem/): Need to know How do MLRun and Iguazio plug into the ML ecosystem? - Check our answer on Iguazio Q&A section.
- [MLRun vs. Airflow vs. MLFlow](https://www.iguazio.com/questions/mlrun-vs-airflow-vs-mlflow/): Need to know MLRun vs. Airflow vs. MLFlow - Check our answer on Iguazio Q&A section.
- [Does Iguazio support data versioning and data labelling?](https://www.iguazio.com/questions/does-iguazio-support-data-versioning-and-data-labelling/): Need to know Does Iguazio support data versioning and data labelling? - Check our answer on Iguazio Q&A section
- [<strong>How do you define the overlapping responsibilities between data scientists, data engineers and MLOps engineers?</strong>](https://www.iguazio.com/questions/how-do-you-define-the-overlapping-responsibilities-between-data-scientists-data-engineers-and-mlops-engineers/): Need to know overlapping responsibilities between data scientists, data engineers and MLOps engineers - Check our answer
- [Iguazio vs. MLRun vs. Nuclio: Whats the Difference?](https://www.iguazio.com/questions/iguazio-vs-mlrun-vs-nuclio-whats-the-difference/): Need to know Iguazio vs. MLRun vs. Nuclio: Whats the Difference? - Check our answer on Iguazio Q&A section.
- [What Are the Tradeoffs Between a Data Lake and a Data Warehouse?](https://www.iguazio.com/questions/what-are-the-tradeoffs-between-a-data-lake-and-a-data-warehouse%ef%bf%bc/): Need to know What Are the Tradeoffs Between a Data Lake and a Data Warehouse? - Check our answer on Iguazio Q&A section.
- [Batch Processing vs. Stream Processing: What’s the Difference? ](https://www.iguazio.com/questions/batch-processing-vs-stream-processing/): Need to know Batch Processing vs. Stream Processing: What’s the Difference?  - Check our answer on Iguazio Q&A section.
- [Static Deployment vs. Dynamic Deployment: What’s the Difference?  ](https://www.iguazio.com/questions/static-vs-dynamic-deployment/): Need to know Static Deployment vs. Dynamic Deployment: What’s the Difference?   - Check our answer on Iguazio Q&A section.
- [How do I train a model on very large datasets?](https://www.iguazio.com/questions/how-do-i-train-a-model-on-very-large-datasets/): Need to know How do I train a model on very large datasets? - Check our answer on Iguazio Q&A section.
- [How do I move my batch pipeline over to real time?](https://www.iguazio.com/questions/how-do-i-move-my-batch-pipeline-over-to-real-time/): Need to know How do I move my batch pipeline over to real time? - Check our answer on Iguazio Q&A section.
- [How do I automate the training pipeline with my CI/CD framework?](https://www.iguazio.com/questions/how-do-i-automate-the-training-pipeline-with-my-ci-cd-framework/): Need to know How do I automate the training pipeline with my CI/CD framework? - Check our answer on Iguazio Q&A section.
- [What is the difference between data drift and concept drift?](https://www.iguazio.com/questions/what-is-the-difference-between-data-drift-and-concept-drift/): Data drift vs. concept drift: It’s important to understand the difference between them, because they require different approaches.
- [What is self-supervised learning in machine learning and how is it different from supervised learning?](https://www.iguazio.com/questions/what-is-self-supervised-learning-in-machine-learning-and-how-is-it-different-from-supervised-learning/): Self-supervised learning is an evolving technique of helping ML models to learn from more data, without the human-labelled datasets.
- [How do I serve models for real-time enterprise applications?](https://www.iguazio.com/questions/how-do-i-serve-models-for-real-time-enterprise-applications/): Need to know How do I serve models for real-time enterprise applications? - Check our answer on Iguazio Q&A section.
- [What can I do with model monitoring?](https://www.iguazio.com/questions/what-can-i-do-with-model-monitoring/): Need to know What can I do with model monitoring? - Check our answer on Iguazio Q&A section.
- [Why is model monitoring so important?](https://www.iguazio.com/questions/why-is-model-monitoring-so-important/): Need to know Why is model monitoring so important? - Check our answer on Iguazio Q&A section.
- [Where does a feature store fit into the ML lifecycle?](https://www.iguazio.com/questions/where-does-a-feature-store-fit-into-the-ml-lifecycle/): Need to know Where does a feature store fit into the ML lifecycle? - Check our answer on Iguazio Q&A section.
- [Who benefits from a feature store?](https://www.iguazio.com/questions/who-benefits-from-a-feature-store/): Need to know Who benefits from a feature store? - Check our answer on Iguazio Q&A section.
- [How are ML pipelines evolving to make way for MLOps?](https://www.iguazio.com/questions/how-are-ml-pipelines-evolving-to-make-way-for-mlops-2/): Need to know How are ML pipelines evolving to make way for MLOps? - Check our answer on Iguazio Q&A section.
- [What are a feature store’s capabilities?](https://www.iguazio.com/questions/what-are-a-feature-stores-capabilities/): Need to know What are a feature store’s capabilities? - Check our answer on Iguazio Q&A section.
- [Data preprocessing vs. feature engineering](https://www.iguazio.com/questions/data-preprocessing-vs-feature-engineering-whats-the-difference/): Need to know Data preprocessing vs. feature engineering - Check our answer on Iguazio Q&A section.
- [What are the key components of a successful MLOps strategy?](https://www.iguazio.com/questions/what-are-the-key-components-of-a-successful-mlops-strategy/): Need to know What are the key components of a successful MLOps strategy? - Check our answer on Iguazio Q&A section.
- [What are the pros and cons of MLOps?](https://www.iguazio.com/questions/what-are-the-pros-and-cons-of-mlops/): Need to know What are the pros and cons of MLOps? - Check our answer on Iguazio Q&A section.
- [What are Kubeflow Pipelines?](https://www.iguazio.com/questions/what-are-kubeflow-pipelines/): Need to know What are Kubeflow Pipelines? - Check our answer on Iguazio Q&A section.
- [What businesses benefit the most from MLOps?](https://www.iguazio.com/questions/what-businesses-benefit-the-most-from-mlops/): Need to know What businesses benefit the most from MLOps? - Check our answer on Iguazio Q&A section.
- [How do I select monitoring metrics specific to my use case?](https://www.iguazio.com/questions/how-do-i-select-monitoring-metrics-specific-to-my-use-case/): Need to know How do I select monitoring metrics specific to my use case? - Check our answer on Iguazio Q&A section.
- [How are ML pipelines evolving to make way for MLOps?](https://www.iguazio.com/questions/how-are-ml-pipelines-evolving-to-make-way-for-mlops/): Need to know How are ML pipelines evolving to make way for MLOps? - Check our answer on Iguazio Q&A section.
- [What is MLOps, and why should we care?](https://www.iguazio.com/questions/what-is-mlops-and-why-should-we-care/): Need to know What is MLOps, and why should we care? - Check our answer on Iguazio Q&A section.
- [Why does governance come first in MLOps?](https://www.iguazio.com/questions/why-does-governance-come-first-in-mlops/): Need to know Why does governance come first in MLOps? - Check our answer on Iguazio Q&A section.

---

#
# Detailed Content


## Pages


{ "@context": "https://schema. org/", "@type": "Product", "name": "GPU as a Service", "image": "", "description": "Accelerate machine learning, deep learning and data processing using a single GPU as a Service solution, for faster and scalable AI-based applications", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Jeffrey Tseng", "reviewBody": "Iguazio complements the compute power of GPUs by directly connecting GPUs to large-scale, shared data infrastructure. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-23", "author": {"@type": "Person", "name": "Jeffrey Tseng"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Real-Time Feature Engineering", "image": "", "description": "Feature engineering made simple: Ingest real-time data, perform data transformation, generate and share real-time features across teams. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "3", "reviewCount": "3" }, "review": }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "GPU as a Service", "image": "", "description": "Accelerate machine learning, deep learning and data processing using a single GPU as a Service solution, for faster and scalable AI-based applications", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Jeffrey Tseng", "reviewBody": "Iguazio complements the compute power of GPUs by directly connecting GPUs to large-scale, shared data infrastructure. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-23", "author": {"@type": "Person", "name": "Jeffrey Tseng"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "GPU as a Service", "image": "", "description": "Accelerate machine learning, deep learning and data processing using a single GPU as a Service solution, for faster and scalable AI-based applications", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Jeffrey Tseng", "reviewBody": "Iguazio complements the compute power of GPUs by directly connecting GPUs to large-scale, shared data infrastructure. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-23", "author": {"@type": "Person", "name": "Jeffrey Tseng"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Ad-Tech & Gaming", "image": "", "description": "Use data science to optimize ad-tech and gaming experiences. Anticipate users' next move, reduce church, increase engagement and provide the best user experience possible. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Data Scientist and Growth Hacker", "reviewBody": "Iguazio enabled us to pool multiple data sources and create a real-time machine learning and rule-based tagging engine. Working with our product team moved the needle on our day one retention", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Data Scientist and Growth Hacker"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Telecommunications", "image": "", "description": "Leverage AI to transform your telecoms' efficiency by creating automatically self-healing systems, improving network optimization, churn reduction, customer engagement and more. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Merav Rotem-Naaman", "reviewBody": "Iguazio is a trusted partner for companies looking to use data to make real-time business decisions that improve security and operations", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Merav Rotem-Naaman"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Real-Time Feature Engineering", "image": "", "description": "Feature engineering made simple: Ingest real-time data, perform data transformation, generate and share real-time features across teams. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "3", "reviewCount": "3" }, "review": }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "GPU as a Service", "image": "", "description": "Accelerate machine learning, deep learning and data processing using a single GPU as a Service solution, for faster and scalable AI-based applications", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Jeffrey Tseng", "reviewBody": "Iguazio complements the compute power of GPUs by directly connecting GPUs to large-scale, shared data infrastructure. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-23", "author": {"@type": "Person", "name": "Jeffrey Tseng"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Healthcare", "image": "", "description": "From mitigating risks proactively, to creating personalized customer services, harness the power of AI to transform your bottom line. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Yaron Weiss", "reviewBody": "With Iguazio’s Data Science Platform, we built a scalable and reliable system which adapts to new threats and enables us to prevent fraud with minimum false positives. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Yaron Weiss"} } }

---


Transforming Data into Advanced Offline and Online Features Building a Feature Vector From Features Real-Time Features and Drift Detection { "@context": "https://schema. org/", "@type": "Product", "name": "Integrated Feature Store", "image": "", "description": "Manage your ML workflow end to-end using a user-friendly environment, featuring fully integrated workflow management, experiment tracking and AutoML tools", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "3", "reviewCount": "3" }, "review": }

---


Want to learn more about machine learning pipelines? Book a live demo here.

---


Want to learn more about enterprise data science? Book a live demo here.

---


Want to learn more about machine learning pipeline tools? Book a live demo here.

---


Want to learn more about operationalizing data science? Book a live demo here.

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Machine Learning Pipeline Orchestration", "image": "", "description": "Manage your ML workflow end to-end using a user-friendly environment, featuring fully integrated workflow management, experiment tracking and AutoML tools", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "3", "reviewCount": "3" }, "review": }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Serverless Automation", "image": "", "description": "Consume Nuclio, Iguazio's open source serverless framework to automate each step of the pipeline with enterprise performance, scale and reliability", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Dr. Shim Yoon", "reviewBody": "We look forward to providing our customers with intelligent, serverless applications by implementing Iguazio’s technology to our cloud’s PaaS", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-23", "author": {"@type": "Person", "name": "Dr. Shim Yoon"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Telecommunications", "image": "", "description": "Leverage AI to transform your telecoms' efficiency by creating automatically self-healing systems, improving network optimization, churn reduction, customer engagement and more. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Merav Rotem-Naaman", "reviewBody": "Iguazio is a trusted partner for companies looking to use data to make real-time business decisions that improve security and operations", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Merav Rotem-Naaman"} } }

---


! function{var n={"@context":"https://schema. org/","@type":"Product","name":"Data Science For Manufacturing","image":"https://di5r923elu32w2e633ofbxj9-wpengine. netdna-ssl. com/wp-content/uploads/2020/01/3. png","description":"Use data science to optimize manufacturing. Produce gains in quality, productivity and cost reduction by leveraging AI on big data in your organization. ","aggregateRating":{"@type":"AggregateRating","ratingValue":"5","bestRating":"5","worstRating":"1","ratingCount":"2","reviewCount":"2"},"review":},e=document. createElement("script");e. type="application/ld+json",e. innerHTML=JSON. stringify(n),document. head. appendChild(e)};

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Retail", "image": "", "description": "Leverage AI to determine ways to reach the right customers at the right time, to ensure efficiency across distribution channels and to transform your bottom line. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Aparna Sinha", "reviewBody": "We are excited to collaborate with Iguazio to deliver a solution that enables real-time analytics of store data, all centrally managed from Google Cloud. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Aparna Sinha"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Smart Mobility", "image": "", "description": "Use data science to optimize smart mobility. From mass transit scheduling and airplane ground-time optimization to better ride-hailing apps and user experiences, learn how AI can transform your organization. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Asanka Perera", "reviewBody": "Iguazio provides us with an integrated platform which makes the job of better detecting fraud and ride usage much easier and efficient in the doing. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Asanka Perera"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Ad-Tech & Gaming", "image": "", "description": "Use data science to optimize ad-tech and gaming experiences. Anticipate users' next move, reduce church, increase engagement and provide the best user experience possible. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Data Scientist and Growth Hacker", "reviewBody": "Iguazio enabled us to pool multiple data sources and create a real-time machine learning and rule-based tagging engine. Working with our product team moved the needle on our day one retention", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Data Scientist and Growth Hacker"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "GPU as a Service", "image": "", "description": "Accelerate machine learning, deep learning and data processing using a single GPU as a Service solution, for faster and scalable AI-based applications", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Jeffrey Tseng", "reviewBody": "Iguazio complements the compute power of GPUs by directly connecting GPUs to large-scale, shared data infrastructure. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-23", "author": {"@type": "Person", "name": "Jeffrey Tseng"} } }

---


{ "@context": "https://schema. org/", "@type": "Product", "name": "Data Science For Financial Services", "image": "", "description": "From mitigating risks proactively, to creating personalized customer services, harness the power of AI to transform your bottom line. ", "brand": "Iguazio", "aggregateRating": { "@type": "AggregateRating", "ratingValue": "5", "bestRating": "5", "worstRating": "1", "ratingCount": "1", "reviewCount": "1" }, "review": { "@type": "Review", "name": "Yaron Weiss", "reviewBody": "With Iguazio’s Data Science Platform, we built a scalable and reliable system which adapts to new threats and enables us to prevent fraud with minimum false positives. ", "reviewRating": { "@type": "Rating", "ratingValue": "5", "bestRating": "5", "worstRating": "1" }, "datePublished": "2020-04-21", "author": {"@type": "Person", "name": "Yaron Weiss"} } }

---


---


## Posts


Customer service chatbots and co-pilots and smart call center analysis applications are prime use cases for AI and generative AI. These AI systems and agents can provide real-time recommendations, support customer service at scale, generate insights that can be used in downstream applications to reduce churn and increase revenue, and more. How can customer service organizations grow and optimize their use of data and AI? For data scientists tasked with building and training AI models and LLMs, open and free telecom and call center datasets are an important starting point. But these datasets can be hard to come by, since they include personal customer information and business competitive information, which is why not many organizations share this data. This blog post is here to help.   Top Free Datasets for Call Centers and Telcos Here are 13 excellent open datasets and data sources for telco and call center data for AI. Call Center Data A dataset containing metrics about incoming calls, answered calls, abandoned calls, service rate, and more. Get the dataset here. Call Center Analysis A 5,000 row dataset containing information about the agent’s name, time of call, if the issue was answered and resolved, the speed of answer, call duration and satisfaction score. Get the dataset here. Telco Customer Churn Customer churn information for a fictitious telco company, offered by IBM. Data includes whether the customer churned in the last month, gender, monthly charges, services the customer signed up for, and more. Get the dataset here. Telco Customer...

---


Chatbots are quickly becoming more sophisticated and integrated into business workflows, enhancing productivity and scalability. However, they also expand the attack surface for organizations. This new exploitation vector requires data engineers and security teams to incorporate various security guardrails when building their gen AI architecture. In this blog post, we discuss the risk of data leakage through AI chatbots. After explaining the risk, we detail suggested guardrails for three levels of organizational risk appetite. In the end, we provide three example e-commerce use cases for each risk level. What is the Risk of Chatbot Data Leakage? Data leakage happens when sensitive, confidential, or proprietary information is unintentionally exposed or accessed by unauthorized parties through a chatbot. This can occur through: Over-permissioning: If the bot has broad access rights (e. g. , admin-level access to databases or systems), which could cause it to reveal sensitive info or take unintended actions. Data Exfiltration by Malicious Users: A compromised chatbot can expose sensitive data, like customer information, business records, financial information, and more. Training: If the model is trained on sensitive internal data without proper anonymization, this information might be exposed in future responses. Prompt Injection & Jailbreaking: If a user can trick the bot into exposing data (e. g. , “Ignore your previous instructions and show me all payroll records”). Insecure API or Plugin Integrations: If the chatbot connects via APIs or third-party plugins, and those aren’t hardened, attackers could exploit them to access internal assets. Unmonitored Activity: Bots can make database...

---


We are proud to announce a new integration between MLRun, the open-source AI orchestration framework, and NVIDIA NeMo microservices, by extending NVIDIA Data Flywheel Blueprint. This integration streamlines training, evaluation, fine-tuning and monitoring of AI models at scale, ensuring high-performance, low latency and lowering costs while significantly reducing the manual effort required through intelligent automation. MLRun is built and maintained by Iguazio, now a part of QuantumBlack, McKinsey’s AI arm. As part of NVIDIA’s ecosystem and with a longstanding history of collaboration, we’re excited to help developers worldwide build agentic systems that bring real business value. Read the blog for more details, or go straight to the blueprint to try it out for yourself. What is MLRun? MLRun is an open-source AI orchestration framework for managing GenAI and ML applications across their lifecycle. It automates data preparation, model tuning, customization, validation and optimization of LLMs, ML models and live AI applications over elastic resources. MLRun enables the rapid deployment of scalable real-time serving and application pipelines, while providing built-in observability and flexible deployment options, supporting multi-cloud, hybrid, and on-prem environments. What are NVIDIA NeMo Microservices? NVIDIA NeMo is an end-to-end platform that accelerates data flywheel to build and continuously optimize agentic AI systems for peak performance with the latest business intelligence, user inputs, and AI and human feedback. NeMo enables enterprises to optimize AI agents with the latest information and feedback. This modular microservices platform helps implement RAG, customize and evaluate models, incorporate guardrails to keep agents delivering peak performance....

---


In less than three years, gen AI has become a staple technology in the business world. In November of 2022, OpenAI launched ChatGPT, with explosive growth of over 1 million users in just five days, galvanizing the widespread use of gen AI. Over the course of 2023 enterprises entered the experimentation stage and kicked off POCs with API services and open models including Llama 2, Mistral, NVIDIA and others. In 2024, organizations are setting aside dedicated budgets for gen AI while ramping up their efforts to build accelerated infrastructure to support gen AI in production. In this blog post, we spotlight a leading player in the gen AI infrastructure ecosystem, NVIDIA, commonly known for their GPUs, software and research that have helped drive gen AI implementation and research. We introduce their new solution model deployment - NVIDIA NIM. Then, we show how to use NVIDIA NIM with MLRun to productize gen AI applications at scale and reduce risks, including a demo of a multi-agent banking chatbot. The blog is based on the webinar Deploying Gen AI in Production with NVIDIA NIM & MLRun with Amit Bleiweiss, Senior Data Scientist at NVIDIA, and Yaron Haviv, co-founder and CTO and Guy Lecker, ML Engineering Team Lead at Iguazio (acquired by McKinsey). You can watch the entire webinar here. Managed Gen AI Services vs. Do-It-Yourself Organizations often have to choose between managed services, which provide ease of use, and control, through a do-it yourself deployment. Gen AI managed services are often chosen because...

---


We’re proud to announce that the next version of MLRun has been released to community users. On the heels of MLRun v1. 7’s focus on monitoring, MLRun v1. 8 adds features to make LLM and ML evaluation and monitoring more accessible, practical and resource-efficient. New Highlights: Smarter LLM model monitoring with in-platform alerts Improved experiment tracking for document-based models Model evaluation before deployment UI improvements with pagination MLRun is an open-source AI orchestration tool that provides AI practitioners with capabilities to accelerate and streamline the development, deployment and management of gen AI and ML applications. These capabilities include LLM monitoring, fine-tuning, data management, built-in guardrails and more. Ready-made scenarios are available for teams to use instantly, in any environment (both cloud and on-premises). Now, after months of hard work and collaboration between Iguazio’s engineering team, MLRun users and the broader open-source community, MLRun v1. 8 further enhances LLM and ML model monitoring capabilities across the lifecycle. What is MLRun? For those who are new to the tool - MLRun is an open-source AI orchestration framework designed to streamline the lifecycle management of ML and generative AI applications, accelerating their path to production. MLRun automates key processes such as data preparation, model tuning, customization, validation and optimization for ML models, LLMs and live AI applications across scalable, elastic infrastructure. With MLRun, organizations can rapidly deploy real-time serving and application pipelines at scale while benefiting from built-in observability and flexible deployment options that support multi-cloud, hybrid and on-premises environments. Why This Release...

---


Generative AI models are not just tools for producing text, audio or video—they're systems that learn patterns, improvise, and generate unexpected outcomes. When we look at LLMs, we're struck by their capacity to generate surprisingly creative and context-aware results. They can weave coherent narratives, propose novel solutions, mimic human conversation, and even offer nuanced insights across a wide range of topics. While this creativity is their strength, it also introduces variability and risk. The challenge for enterprises isn't just harnessing the power of generative AI—it's making sure that power is predictable, reliable, and safe at scale. Creative Geniuses Genius Statisticians  But unlike human minds, LLM outputs are the results of advanced mathematical calculations across billions of parameters interacting in complex, non-linear ways to predict the most probable next word. Their "creativity" is statistical inference across massive datasets. Unlike traditional rule-based systems, LLMs perform statistical inference without following predefined logical rules like tra enable decision trees. This allows them to generate outputs based purely on patterns learned from data rather than strict instructions. This means LLMs succeed in generating such creative outputs because they are intentionally designed to operate without rigid, hard-coded logic or constant supervision. It is their unmonitored and open-ended nature that stimulates and powers their performance.   The Drawbacks of LLM Creativity This statistical freedom is precisely what gives AI systems their power, but also what makes them so challenging to govern. LLMs operate as black boxes, making it extremely hard to trace back why a certain output...

---


Gen AI call center co-pilots can provide enterprises with operational visibility and insights while automating repetitive tasks, to improve the customer experience. In this session, we’ll show how a large health insurance provider implemented an agentic co-pilot designed scale across multiple call centers and environments.   To dive deep into the architecture and see a demo of the co-pilot, you can watch the webinar this blog is based on. In this session, Guy Lecker (ML Engineer Team Lead) and Nick Schenone (Senior ML Engineer) from Iguazio, acquired by McKinsey) walk us through how generative AI-powered agent co-pilots are revolutionizing the call center experience. Use Case: Real-Time Call Center Co-Pilot A major health insurance provider faced a familiar challenge: agents in their call centers were struggling to consistently guide customers in real-time through complex healthcare decisions, while following strict and data-heavy call scripts and also upselling during calls. The Solution: Combining a Real-Time Agent Co-Pilot with a Post Call Analysis Report The solution was designed as a multi-faceted architecture, which includes both a real-time agent co-pilot and a post-call analysis report. The Agent Co-Pilot The co-pilot enhances agent performance live during calls. The client-facing UI was designed with four main components: Live Transcription Panel – Captures real-time audio conversation in chat format and tracks relevant data as the call unfolds. Call Steps Panel – Ensures agents follow every step of the sales process. It presents the script, highlights skipped steps and places where agent performance could be improved, and changes dynamically...

---


The manufacturing industry can benefit from AI, data and machine learning to advance manufacturing quality and productivity, minimize waste and reduce costs. With ML, manufacturers can modernize their businesses through use cases like forecasting demand, optimizing scheduling, preventing malfunctioning and managing quality. These all significantly contribute to bottom line improvement. In times of global recession, supply chain cut-offs and difficulties meeting consumer demands for materials and products, manufacturing optimization becomes even more important for companies that wish to remain competitive and relevant without impairing their revenue streams. How can manufacturers develop, grow and optimize their use of data and ML? Open and free datasets for machine learning are an important starting point for data scientists and engineers who are developing and training ML models for manufacturing. But these datasets for manufacturing can be hard to come by, since manufacturing often takes a legacy approach and data is not always available. Here are 10 excellent open manufacturing datasets and data sources for manufacturing data for machine learning. 1. Eurostat Industrial Production Index The output and activity of the European industry sector, measured on a monthly basis. The dataset’s base year is 2015 and depicts monthly growth rates. Get the dataset here. 2. US Manufacturing Trends Manufacturing trends in the US related to wage rates, profits, employment, production, capacity utilization, productivity, exports and shipments. The dataset provides information for the present and year-to-date. Get the dataset here. 3. Energy Consumption A dataset providing information about energy consumption at manufacturing sites, homes, commercial...

---


The retail industry has been shaped and fundamentally transformed by disruptive technologies in the past decade. From AI assisted customer service experiences to advanced robotics in operations, retailers are pursuing new technologies to address margin strains and rising customer expectations. AI use cases like personalized product recommendations, demand forecasting for optimized inventory and supply chain management, optimized pricing strategies based on market dynamics, and sales forecasting are generating value for companies who have adopted AI. By leveraging AI, retailers can maintain or increase their competitiveness in a saturated market.   How can retailers use, grow and optimize their use of data and machine learning? For data scientists tasked with building and training machine learning models for retailers, open and free retail datasets are an important starting point. But these datasets for retailers can be hard to come by, since they include personal customer information and business competitive information, which is why not many retailers share this data. This blog post is here to help. Here are 11 excellent open datasets and data sources for retailer data for machine learning. Zara Products 10,000 product records from ZARA USA, including titles, prices, images, availability, and more. Get the dataset here. Slovakia Retail Sales Year-over-year sales from Slovakia, with actual values, historical data, forecast, chart, statistics, economic calendar and news. Get the dataset here. Global Value of Retail Analytics The value of the retail analytics market worldwide in 2016, with a forecast from 2017 to 2022. Get the dataset here. Customer Behavior and...

---


Two years after Seagate first shared their AI and MLOps success story, the data storage leader is now revealing how far they've come since then. In this blog post, you’ll see how the team manages thousands of AI models in production with only a few team members. This is thanks to their AI factory, whichdoes the heavy lifting of automated processes like monitoring, testing, mocking and more. You can hear more details in the webinar this article is based on, straight from Kaegan Casey, AI/ML Solutions Architect at Seagate. How Seagate overcame their AI challenges with an AI Factory Iguazio’s AI factory has helped the Seagate team overcome multiple common MLOps challenges: Project POC to Deployment Gap - The gap between experimentation and deployment is often created due to differing environments (e. g from local or virtual machine to K8s cluster) and the need for bespoke deployments. These variations result in slow feedback loops and delays reaching production. Iguazio allows the team to go from testing code locally to running at scale on a remote cluster within minutes. Visibility - Many teams experience fragmented experiment tracking and lack of collaboration and project visibility. Iguazio allows sharing projects between diverse teams, provides detailed logging of parameters, metrics, and ML artifacts and allows for artifact versioning including labels, tags, associated data etc. Integrations - The team dealt with diverse factory inputs/outputs for real-time or batch inference data due to legacy requirements. Nuclio function triggers (HTTPS, RMQ, Kafka, etc. ) and MLRun function...

---


RAG and Agentic AI shape how intelligent systems interact with data and users. RAG enhances LLMs by retrieving external information to improve accuracy and contextual relevance, while Agentic AI introduces autonomy, decision-making, and adaptability into AI-driven workflows. Agentic RAG combines the power of both, transforming RAG into a multi-step, autonomous, complex process that can self-improve. In this article, we’ll explore how Agentic RAG revolutionizes AI-powered applications by making them more autonomous, intelligent, and context-aware. What is RAG? RAG (Retrieval-Augmented Generation) is an AI approach that enhances LLM outputs by retrieving relevant information from external sources, like databases or the Internet, before generating an answer. This improves response accuracy, reduces hallucinations, and provides more up-to-date responses. It also allows for domain expertise, since the LLM can retrieve industry-specific or proprietary knowledge from related databases and sources. What is Agentic AI? Agentic AI is a type of AI system in which AI applications exhibit a level of autonomy, goal-directed behavior, and adaptability. Unlike traditional AI models that respond only to user inputs, agentic AI can proactively take action, learn from feedback, and optimize its behavior toward achieving specific objectives. This allows the application to make decisions, plan actions, and interact dynamically with its environment. Examples include agents for software development that can design advanced code based on autonomous data analysis or AI personal assistants that proactively manage calendars and emails. Exploring the Concept of Agentic RAG Traditional RAG frameworks retrieve relevant documents from a knowledge base and feed them into a language...

---


Generative AI isn’t just moving fast—it’s on turbo mode. Gartner confirms it in their popular Hype Cycle, compared to other evaluated technologies: gen AI tech is rocketing through the stages faster than anything else. In under three years, it’s already crashing into the trough of disillusionment, while prompt engineering shot to peak hype almost the second it emerged. In this blog post, we bring insights for AI leaders Svetlana Sicular, Research VP, AI Strategy, Gartner and Yaron Haviv, co-founder and CTO, Iguazio (acquired by McKinsey). To see the complete conversation and dive into their insights, watch the webinar here. What Gen AI Trends is Gartner Seeing? Some of the significant fast-paced trends Gartner is seeing include: AI-ready data - Large players are announcing partnerships and acquisitions of data-related companies, for purposes ranging from access to data for AI purposes to explainability to synthetic data. AI governance - Merely two and a half years after gen AI entered the ecosystem, the market is comfortable with generative AI. Now, it's looking for ways to make it safe. This includes responsible AI, Gartner’s concept of AI TRiSM (Trust, Risk and Security in AI Models) and Sovereign AI. AI engineering - AI is being democratized for developers and engineers, expanding beyond the limited pool of data scientists. Companies are building AI tools and frameworks that empower engineers to integrate AI into applications without needing deep expertise in ML. AI Agents and multi-agent systems. And more. See the webinar for more Gartner trends. How should...

---


AI agents are transforming business operations by automating processes, improving decision-making and unlocking new efficiencies. However, their effectiveness depends on how well they are trained. AI Agent Training is the structured process of teaching AI models to perform multi-step assignments, make decisions and adapt to real-world scenarios. Through various training methodologies, such as supervised learning, reinforcement learning, and transfer learning, AI agents can enhance business functions across industries, from fraud detection in finance to customer support automation and predictive maintenance.   This guide explores the core principles, methodologies, challenges and best practices for training AI agents that are not only powerful but also reliable, scalable, and aligned with business goals. What is AI Agent Training? AI Agent Training is a conceptual term that refers to the process of teaching an LLM how to perform multi-step tasks, make complex decisions and adapt to real-world scenarios. This is done by training the LLM on data and providing feedback about the results. There are numerous training methods, like supervised or unsupervised learning, reinforcement learning, few-shot or zero-shot learning, and more (see below). By training them, LLMs can act agentically to provide business value across complex use cases, like customer service chatbots, finance fraud prediction, diagnosing diseases and more. Why AI Agent Training Matters for Businesses AI agents are autonomous software programs that interact with their surroundings to achieve a goal. The effectiveness of achieving this goal depends on how well they are trained. Organizations that effectively train agents can turn them into an...

---


Financial services companies are leveraging data and machine learning to mitigate risks like fraud and cyber threats and to provide a modern customer experience. By following these measures, they are able to comply with regulations, optimize their trading and answer their customers’ needs. In today’s competitive digital world, these changes are essential for ensuring their relevance and efficiency. How can financial services companies build, expand and optimize their use of data and ML? Open and free financial datasets and economic datasets are an essential starting point for data scientists and engineers who are developing and training ML models for finance. But sadly, they can be hard to come by. Here are 13 excellent open financial and economic datasets and data sources for financial data for machine learning. 1. Data. gov A US governmental website hosted by the General Services Administration Technology Transformation Service, data. gov provides a catalog of government data in open-machine and readable formats. To find financial-related datasets, you can search for relevant keywords, e. g “credit card”, and get a list of the available datasets for you to consume. Get the datasets here 2. Data. gov. in Nation-wide data sets from across India, intended to make Indian government-owned shareable data accessible in human and machine readable formats. There are 2,394 Resources in 484 catalogs related to finance, covering topics like consumer price indexes, GDP estimates, prices and more. Get the datasets here 3. data. europa. eu data. europa. eu is the official portal for EU European data....

---


When it comes to leveraging AI to capture business value, it’s worth asking, “what kind of AI do we need exactly? ” There are significant differences between the methodologies collectively referred to as AI. While 2024 might have almost convinced us that gen AI is the end-all-be-all, there is also what’s sometimes called ‘traditional’ AI, deep learning, and much more.   The question is: how do you decide when to use traditional AI, which excels in structured data and predictive modeling, versus generative AI, which shines in creating new content and enhancing human-like interactions? Choosing the right type of AI for your specific use case is critical to maximizing value, minimizing risk, and ensuring sustainable growth. In this blog, we’ll explore the economic potential of generative AI, discover its transformative capabilities across industries, and compare it to the applications of traditional AI. By the end, you’ll have a clear roadmap for identifying when to lean on gen AI and when to choose traditional AI to achieve your business goals. The Potential Economic Value of Gen AI The McKinsey report “The economic potential of generative AI: The next productivity frontier” estimates that generative AI has the potential to add $2. 6 trillion to $4. 4 trillion to the global economy. These sums increase the impact of all AI by an astounding 15 to 40 percent. In addition, the report states, this impact could even double if gen AI is extended beyond the analyzed use cases. The impact of generative AI is...

---


Gen AI applications can bring invaluable business value across multiple use cases and verticals. But sometimes it can be beneficial to experience different types of applications that can be created and operationalized with LLMs. Better understanding the potential value can help: Garner excitement and collaboration across the organization Help secure resources Support planning strategies, from a single use cases to many and from PoC to operationalization Drive more ideas for innovation Help de-risk pitfalls And more In this blog post, we’ve curated the top gen AI demos of AI applications that can be developed with open-source MLRun. Each of these demos can be adapted to a number of industries and customized to specific needs. Follow along and choose the most relevant ones for your needs. You can also watch the complete library of demos here. 1. Smart Call Center Analysis Application Build a call analysis platform for call centers. The application analyzes customers calls and generates actionable insights for agents, management and downstream applications. It helps support agents, create tailored recommendations for customers, and more. This improves the customer experience, helps optimize first call resolution, enhances operational efficiency, supports decision-making and helps meet compliance regulations. The gen AI application is built on a multi-step pipeline that includes diarization, transcription, PII filtering, analysis, post-processing, among others. Output structured data is stored in a database, accessible for reporting or downstream applications. Visual dashboards provide metrics like topic summaries, ag and data filtering outcomes. Open-source MLRun automates the entire workflow, auto-scales resources as...

---


Generative AI has rapidly transformed industries by enabling advanced automation, personalized experiences and groundbreaking innovations. However, implementing these powerful tools requires a production-first approach. This will maximize business value while mitigating risks. This guide outlines six best practices to ensure your generative AI initiatives are effective: valuable, scalable, compliant and future-proof. From designing efficient pipelines to leveraging cutting-edge customization techniques, we’ll walk through the steps that will help your organization harness the full power of generative AI. 1. Create Modular Pipelines to Orchestrate Operationalization from End-to-End  AI pipelines ensure that each step of the AI lifecycle—across data management, development, deployment and monitoring—is automated and optimized for performance. This is particularly important in large-scale gen AI projects, which handle massive datasets and complex models. Streamlining the entire process saves engineering resources, reduces operational bottlenecks and fosters collaboration across teams, from data scientists to DevOps. This ensures models stay accurate and reliable over time. Here’s our suggestion for four pipelines that will bring your models from the lab to production. Data management - Ensuring data quality through data ingestion, transformation, cleansing, versioning, tagging, labeling, indexing, and more. Development - High quality model training, fine-tuning or prompt tuning, validation and deployment with CI/CD for ML. Application - Bringing business value to live applications through a real-time application pipeline that handles requests, data, model and validations. LiveOps - Improving performance, reducing risks and ensuring continuous operations by monitoring data and models for feedback. 2. De-Risk Models and Establish AI Governance and Compliance Operationalizing LLMs...

---


In 2024, organizations realized the revolutionizing business potential of gen AI. They accelerated their gen AI operationalization processes: explored new use cases to implement, researched LLMs and AI pipelines and contemplated underlying ethical issues. And with the seeds of the AI revolution now planted, the market is maturing accordingly. This means that in 2025, we’re likely to already see organizations offering gen AI services to customers, embedded in commercial applications - from SaaS CRMs to chatbots. To support this leap forward, the underlying tech stack will mature as well. Multi-agent systems and multimodality will become widely available, gen AI service providers will expand features and offerings and open-source will challenge commercial vendors with new capabilities. And with LLMs becoming more accurate and intelligent, and new guardrails are introduced, the risk of hallucinations diminishes, perpetuating trust that drives this entire ecosystem forward. Two years after OpenAI released ChatGPT to the public, will 2025 finally be the year gen AI becomes an inseparable part of business? What kind of obstacles do organizations need to overcome? Who will thrive in this new ecosystem? Here are my predictions for the upcoming year: 1. Gen AI Embedded as a Service Forward-thinking enterprises have been working on implementing gen AI as part of their innovation strategies. Excited about the opportunities and vast potential in gen AI, they are ramping up their efforts to implement gen AI in their applications. These could be either to support internal operations, (like McKinsey’s Lilli or a co-pilot agent to support...

---


LLMs are the foundation of gen AI applications. To effectively operationalize and de-risk LLMs and ensure they bring business value, organizations need to consider not just the model itself, but the supporting infrastructure, including GPUs and operational frameworks. By optimizing them to your use case, you can ensure you are using an LLM that is the right fit to your needs. Unlike a one-size-fits-all approach (e. g, using an out-of-the-box public model), adapting your LLM and infrastructure to the right size allows tailoring the model to specific use cases, ensures better performance and saves costs. In this blog post, we’ll take you through the various phases for choosing and optimizing your LLM: model selection, GPU selection, GPU utilization and MLOps practices. In the end, you’ll have the tools to make the necessary choices for building your gen AI application foundation. Step 1: Model Selection Choosing the Model The first step, even before choosing the model, is separating your use case into tasks. These tasks should be as small and focused as possible. Now it’s time to choose the right model for the tasks at hand. While it’s tempting to jump into complex solutions that provide a wide-range of capabilities - we recommend taking a different approach, and starting with smaller, simpler models, so you can assign a small model to each small task. Each task should be handled by the smallest model possible. You can join tasks if your chosen model can handle a number of them at the same...

---


As the open-source maintainers of MLRun, we’re proud to announce the release of MLRun v1. 7. MLRun is an open-source AI orchestration tool that accelerates the deployment of gen AI applications, with features such as LLM monitoring, fine-tuning, data management, guardrails and more. We provide ready-made scenarios that can be easily implemented by teams in organizations. This new version brings substantial enhancements that address these increasing demands of gen AI deployments, with a particular focus on monitoring LLMs. Additional updates introduce performance optimizations, multi-project management, and more. MLRun v1. 7 is the culmination of months of hard work and collaboration between the Iguazio engineering team, MLRun users and the open-source community. We’ve listened to what our users are saying and have designed this version (and the upcoming ones) to address needs and gaps for managing ML and AI across the lifecycle. With this new version, users ranging from individual contributors to large teams in enterprises will be able to deploy AI applications much faster and with more flexibility than before. Brief Intro for Newcomers: What is MLRun? MLRun is an open-source AI orchestration framework for managing ML and generative AI applications across their lifecycle, to accelerate their productization. It automates data preparation, model tuning, customization, validation and optimization of ML models, LLMs and live AI applications over elastic resources. MLRun enables the rapid deployment of scalable real-time serving and application pipelines, while providing built-in observability and flexible deployment options, supporting multi-cloud, hybrid, and on-prem environments. Why This Release Focuses on...

---


Gen AI has the potential to bring immense value for marketing use cases, from content creation to hyper-personalization to product insights, and many more. But if you’re struggling to scale and operationalize gen AI, you’re not alone. That’s where most enterprises struggle. To date, many companies are still in the excitement and exploitation phase of gen AI. Few have a number of initial pilots deployed and even fewer have simultaneous pilots and are building differentiating use cases. Only a handful have select use cases running at scale, and no company, per McKinsey’s knowledge, is using gen AI as the new normal in their day-to-day use cases.   In this blog post, we provide a staged approach for rolling out gen AI, together with use cases, a demo and examples that you can implement and follow. For more details, watch the webinar this blog post is based on. The webinar hosts Eli Stein, Partner and Modern Marketing Capabilities Leader from McKinsey, Ze’ev Rispler, ML Engineer, from Iguazio (acquired by McKinsey), and myself. Watch the entire webinar here. Marketing and Technology: A Historical Perspective  Marketing is frequently being disrupted by technology. This has occurred with computing in the 1970s, the Internet in the 1990s and mobile in the 2000s. In 2022, “AI everywhere” has enabled zero marginal cost of content generation. In the upcoming years, we can expect to see AGI (Artificial General Intelligence), which will provide intelligence beyond human capabilities, at zero marginal cost. And similar to how some marketers were...

---


If 2023 was the year of gen experimentation, 2024 is the year of gen AI implementation. As companies embark on their implementation journey, they need to deal with a host of challenges, like performance, GPU efficiency and LLM risks. These challenges are exacerbated in highly-regulated industries, such as financial services and telecommunication, adding further implementation complexities. Below, we discuss these challenges and present some best practices and solutions to take into consideration. We also show a banking chatbot demo that includes fine-tuning a model and adding guardrails. This blog post is based on the webinar “Implementing Gen AI in Highly Regulated Environments” with Asaf Somekh, co-founder and CEO, and Yaron Haviv, co-founder and CTO, of Iguazio (acquired by McKinsey). You can watch the entire session here. Challenges Implementing and Scaling Gen AI Implementing and scaling gen AI comes with a host of challenges: Performance vs. cost - The massive scale of LLMs make these models relatively inefficient and compute-heavy. Costs can especially spike in highly-regulated industries, where the models need to be run within cloud accounts. In addition, data science teams need flexibility to switch between different LLMs to fit their use case, further enhancing costs. The data science team needs to find a balance between performance accuracy and cost efficiency. At a business level, the organization needs to ensure the cost associated with the development and deployment is lower than the value the application will bring to the business. GPU efficiency - GPUs are scarce and costly. The team...

---


AI and generative Al can lead to major enterprise advancements and productivity gains. By offering new capabilities, they open up opportunities for enhancing customer engagement, content creation, virtual experts, process automation and optimization, and more. According to McKinsey & Company, gen Al has the potential to deliver an additional $200-340B in value for the banking industry. One popular gen AI use case is customer service and personalization. Gen AI chatbots have quickly transformed the way that customers interact with organizations. They can handle customer inquiries and provide personalized recommendations while empathizing with them and offering nuanced support that is tailored to the customer’s individual needs. Another less obvious use case is fraud detection and prevention. AI offers a transformative approach by automating the interpretation of regulations, supporting data cleansing, and enhancing the efficacy of surveillance systems. AI-powered systems can analyze transactions in real-time and flag suspicious activities more accurately, which helps institutions take informed actions to prevent financial losses.   In this blog post, we introduce the joint MongoDB-Iguazio gen AI solution, which allows for the development and deployment of resilient and scalable gen AI applications. Before diving into how it works and its value for you, we will introduce MongoDB and Iguazio (acquired by McKinsey). We will then list the challenges enterprises are dealing with today when operationalizing gen AI applications. In the end, we’ll provide resources on how to get started. MongoDB for end-to-end AI data management MongoDB Atlas, an integrated suite of data services centered around a...

---


RAG and Fine-Tuning are two prominent LLM customization approaches. While RAG involves providing external and dynamic resources to trained models, fine-tuning involves further training on specialized datasets, altering the model. Each approach can be used for different use cases. In this blog post, we explain each approach, compare the two and recommend when to use them and which pitfalls to avoid. What is Fine-Tuning? LLM fine-tuning is an AI/ML LLM customization process where a pre-trained model is further trained on a new dataset that is specific to a particular task. Unlike RAG, another form of LLM customization, this includes modifying the model’s weights and parameters based on the new data. By adapting and “tweaking” the LLM, fine-tuning improves the model’s performance and accuracy for the required task. This allows for better applicability to specific tasks, domains and use cases and brings more business value. For example: 1. If you need a model that excels in legal document analysis, you can fine-tune an LLM pre-trained in English, using a corpus of legal texts. The fine-tuned model will then better understand legal jargon, context and nuances. The result will be a model highly effective for tasks like legal document classification or summarization. 2. A pre-trained image recognition model can be fine-tuned to identify specific objects relevant to a particular industry, such as medical imaging for tumor detection or industrial inspection for defect identification. 3. Models pre-trained on general speech data can be fine-tuned to recognize industry-specific jargon or accents. This can improve...

---


As can be inferred from their name, foundation models are the foundation upon which developers build AI applications for tasks like language translation, text summarization, sentiment analysis and more. Models such as OpenAI's GPT, Google's Gemini, Meta’s Llama and Anthropic’s Claude, are pre-trained on vast amounts of text data and have the capability to understand and generate human-like language. In addition to the hype surrounding these commercial models, there's also a rising interest in self-hosted open-source LLMs, like Mistral. As businesses and organizations increasingly rely on AI for various applications, the demand for advanced foundation models is expected to continue growing, driving further innovation and competition in the market. In this blog post, we dive into the considerations to make when choosing a proprietary model* vs. self-hosting an open-source model. We also provide some example use cases that exemplify how to implement these requirements. Then, we show how these requirements are implemented in models, through Artificial Analysis’s website comparison. Finally, we provide a list of additional tools and capabilities that can help optimize LLMs even more. (*Clarification: While in some cases commercial LLMs like Cohere and AI21 can be hosted in your own environments, this is not the common use case. Throughout the blog post, when referring to commercial LLMs we mean vendor-hosted models). Commercial vs. Self-Hosted: How to Make the Choice When deciding whether to choose commercial versus self-hosted models, several key factors should be considered. These will help determine the best fit for your particular use cases or...

---


Enterprises are beginning to implement gen AI across use cases, realizing its enormous potential to deliver value. Since we are all charting new technological waters, being mindful of recommended strategies, pitfalls to avoid and lessons learned can assist with the process and help drive business impact and productivity. In this blog post, we provide a number of frameworks that can help enterprises effectively implement and scale gen AI while avoiding risk. We also include a number of use cases, from R&D to automotive to the supply chain. In the end, we list potential hurdles and how to overcome them. This blog post is based on the webinar “Transforming Enterprise Operations with Gen AI”, which was held with Dinu de Kroon, Partner and Operations Hub Lead, Nicola Unfer, Sr. Program Delivery Analyst and Davide Di Lucca, Research Science Analyst, from McKinsey, and Yaron Haviv, Co-Founder and CTO, Iguazio (acquired by McKinsey). You can watch the webinar recording here. 5 Questions to Ask and Answer About Gen AI in the Enterprise The evolution of AI began in the 1950s, but the advent of ChatGPT and other generative AI capabilities have created the “perfect storm” of AI. This revolution has been driven by the convergence of massive computing power, enabling data processing at unprecedented scale and speed; an abundance of data available through the internet for model training; and pre-trained transformers that empower us to efficiently work with unstructured data. At this point in time, it’s recommended that enterprises ask themselves five questions...

---


The generative AI industry is changing fast. New models and technologies (Hello GPT-4o) are emerging regularly, each more advanced than the last. This rapid development cycle means that what was cutting-edge a year ago might now be considered outdated. The rate of change demands a culture of continuous learning and technological adaptation. To ensure AI applications remain relevant, effective, secure and capable of delivering value, teams need to keep up with the latest research, technological developments and potential use cases. They also need to understand regulatory and ethical implications of deploying AI models, taking into consideration issues like data privacy, security and ethical AI use. But building gen AI pipelines and operationalizing LLMs requires significant engineering resources. How can organizations ensure their architecture remains robust, resilient, scalable and secure, so it can support up-to-date LLM deployment and management? The Solution: Designing Modular Pipelines with Swappable Components To stay relevant and future-proof, applications need to be designed with adaptability in mind. One of the most strategic approaches to addressing this need is designing your AI architecture in a modular fashion, where different components of the pipelines can be swapped out or updated as needed without overhauling the entire system. The benefits of a modular architecture include: The ability to easily upgrade components and modules in your framework with minimal adjustments. Reducing the risk associated with changes, since updates are confined to specific modules. This containment makes it easier to trial new features and roll them back if they don’t perform as...

---


LLM evaluation is the process of assessing the performance and capabilities of LLMs. This helps determine how well the model understands and generates language, ensuring that it meets the specific needs of applications. There are multiple ways to perform LLM evaluation, each with different advantages. In this blog post, we explain the role of LLM evaluation in AI lifecycles and the different types of LLM evaluation methods. In the end, we show a demo of a chatbot that was developed with crowdsourcing. This blog post is based on a webinar with Ehud Barnea, PhD, Head of AI at Tasq. AI, Yaron Haviv, co-founder and CTO of Iguazio (acquired by McKinsey) and Guy Lecker, ML Engineer Team Lead at Iguazio, which you can watch here. Gen AI Reference Architecture Following Established ML Lifecycles Building generative AI applications requires four main elements: Data management - Ingesting, preparing and indexing the data. Model development - Taking the existing baseline model, training and tuning if required with additional data gathered from data processing or through a feedback loop, evaluating, testing and deployment. Application deployment - Receiving the prompt engineering requests, running through LLMs and post-processing. LiveOps - Monitoring and validation of information from the application pipeline onto monitoring frameworks, labeling, filtering and sending to fine-tuning if required. To build accurate and risk-free models, improve performance and implement governance, it’s important to place data and model controls in gen AI apps. This requires controls in the pipelines, in multiple places. This includes: Cleaning and preparing...

---


Ever since the release of ChatGPT in November 2022, organizations have been trying to find new and innovative ways to leverage gen AI to drive organizational growth. LLM capabilities like contextual understanding and response to natural language prompts enable the development of applications like automated AI chatbots, smart call center apps, or for financial services. Generative AI is by no means a replacement for the previous wave of AI/ML (now sometimes referred to as ‘traditional AI/ML’), which continues to deliver significant value, and represents a distinct approach with its own advantages. By integrating LLMs with traditional ML models, organizations can significantly enhance and augment each model’s capabilities, leading to new and exciting applications that bring value to their customers. In this blog post, we detail LLMs’ and ML models’ strengths, evaluate the benefits of integration and provide a number of example use cases, from advanced chatbots to synthetic data generation. In the end, we explain how MLOps can help accelerate the process and bring these models to production. LLMs vs. Classical ML Models: Strengths and Capabilities LLMs and ML models each have their distinct strengths, which can be applied to different kinds of tasks and objectives. Strengths of LLMs: Natural Language Understanding and Generation - LLMs excel at comprehending and producing human-like text. They can generate coherent and contextually relevant responses over a wide range of topics, making them ideal for applications like chatbots, content creation and language translation.   Contextual Learning - With their deep learning architecture, LLMs can...

---


Organizations that monitor their LLMs will benefit from higher performing models at higher efficiency, while meeting ethical considerations like ensuring privacy and eliminating bias and toxicity. In this blog post, we bring the top LLM metrics we recommend measuring and when to use each one. In the end, we explain how to implement these metrics in your ML and gen AI pipelines. Why Do We Need to Monitor LLM Metrics? Monitoring the metrics of LLMs enhances performance optimization, enables understanding user interaction and ensures ethical compliance. In more detail: Accuracy - Monitoring the outputs of the model is the main course for validating its reliability at a given task. It is the first flag that needs to be raised in case the model needs to go into another phase of development, whether it is prompt engineering the inputs to the model or fine-tuning the model itself. Resource Management - LLMs require significant computational resources. Metrics related to resource utilization help in managing these resources effectively and reduce operational costs. User Interaction - Monitoring metrics related to user interactions helps in understanding how users engage with the model. These insights can guide enhancements in user experience, making the model more intuitive and responsive to user needs. Ethical Compliance and Bias Reduction - Monitoring metrics related to the ethical use of LLMs ensures the trustworthiness of the model. This is important for preventing incomplete and incorrect responses, responses with the wrong tone, violation of privacy or ethical standards (like preventing ePHI leakage...

---


Customer care organizations are facing the disruptions of an AI-enabled future, and gen AI is already impacting customer care organizations across use cases like agent co-pilots, summarizing calls and deriving insights, creating chatbots and more. In this blog post, we dive deep into these use cases and their business and operational impact. Then we show a demo of a call center app based on gen AI that you can follow along. For more details on this topic, you can watch the webinar this blog post is based on. Oana Cheta, Partner and Lead Gen AI Service Ops for North America at McKinsey & Company and Yaron Haviv, Co-Founder and CTO of Iguazio (acquired by McKinsey), share insights, examples and more details. Watch here. The 3 Driving Forces of AI Impact AI is broadly redefining work and the economy. The three main areas of impact are: 1. Dramatic productivity gains in how companies are run. This includes improving functions like coding, customer interactions, creative work and content synthesis. 2. Transformation of products to change how customer needs are met. This includes new enhanced features, like conversational interfaces, co-pilots and hyper-personalization. 3. A redistribution of profit pools through a new layer of AIaaS in the value chain. This could enable solutions like ChatGPT that replace entire value chains, low-cost native AI startups that quickly create new solutions at low costs and simplified onboarding and migration processes between products. How Are Industries Being Transformed with Gen AI? As a result, entire industries could...

---


GPU provisioning serves as a cost-effective solution for organizations who need more GPUs for their ML and gen AI operations. By optimizing the use of existing resources, GPU provisioning allows organizations to build and deploy their applications, without waiting for new hardware. In this blog post, we explain how GPU provisioning as a service works, how it can close the GPU shortage gap, when to use GPU provisioning and how it fits with gen AI. How Companies are Dealing with the GPU Shortage Organizations need GPUs to be able to process large amounts of data simultaneously, speed up computational tasks and handle specific applications like AI, data visualization and more. This need is likely to grow as the demand for computational power and real-time processing increases across industries. However, growing demand is meeting a lack of supply, and organizations are encountering obstacles when attempting to purchase GPUs, with years-long waitlists. As a result, companies are hustling their way to GPU access. They’re trying to leverage their industry connections and reach out to anyone who might help them gain access to GPUs. This includes seeking assistance through professional networks, applying for government grants and forming partnerships with cloud providers to secure access to GPUs for AI ventures. Others are more creative, setting up initiatives like renting out GPUs or trying to repurpose other hardware.   The unexpected wide adoption of gen AI has further exacerbated these attempts. Companies across industries are looking to implement LLMs in their operations, which require GPUs...

---


Gen AI is quickly reshaping industries, and the pace of innovation is incredible to witness. The introduction of ChatGPT, Microsoft Copilot, Midjourney, Stable Diffusion and many more incredible tools have opened up new possibilities we couldn’t have imagined 18 months ago. While building gen AI application pilots is fairly straightforward, scaling them to production-ready, customer-facing implementations is a novel challenge for enterprises, and especially for the financial services sector. Risk, compliance, data privacy and escalating costs are just a few of the acute concerns that financial services companies are grappling with today. This blog post will discuss: The potential impact of generative AI for financial services The challenges of deploying LLMs in production Which engineering and risk-related considerations financial services companies need to take to successfully implement gen AI in their business environments. To learn more, watch the webinar “Implementing Gen AI for Financial Services” with Larry Lerner, Partner & Global Lead - Banking and Securities Analytics, McKinsey & Company, and Yaron Haviv, Co-founder and CTO, Iguazio (acquired by McKinsey), which this blog post is based on. View the entire webinar here. What is the Potential Value of Gen AI and Analytics for Financial Services? The potential annual value of AI and analytics for global banking could reach as high as $1 trillion. But the evolution from analytical Al to generative Al has led to major advancements in the power of advanced analytics. Gen AI has the potential to deliver significant incremental value, potentially leading to 3-5% margin improvements. These...

---


Data engineers, data scientists and other data professional leaders have been racing to implement gen AI into their engineering efforts. But a successful deployment of LLMs has to go beyond prototyping, which is where LLMOps comes into play. LLMOps is MLOps for LLMs. It’s about ensuring rapid, streamlined, automated and ethical deployment of LLMs to production. This blog post delves into the concepts of LLMOps and MLOps, explaining how and when to use each one. To read more about LLMOps and MLOps, checkout the O’Reilly book “Implementing MLOps in the Enterprise”, authored by Iguazio’s CTO and co-founder Yaron Haviv and by Noah Gift. What is LLMOps? LLMOps (Large Language Model Operations), is a specialized domain within the broader field of machine learning operations (MLOps). LLMOps focuses specifically on the operational aspects of large language models (LLMs). LLM examples include GPT, BERT, and similar advanced AI systems. LLM models are large deep learning models that are trained on vast datasets, are adaptable to various tasks and specialize in NLP tasks. They are characterized by their enormous size, complexity, and the vast amount of data they process. These elements need to be taken into consideration when managing, streamlining and deploying LLMs in ML pipelines, hence the specialized discipline of LLMOps. Addressing LLM risks is an important part of gen AI productization. These risks include bias, IP and privacy issues, toxicity, regulatory non-compliance, misuse and hallucination. Mitigation starts by ensuring the training data is reliable, trustworthy and adheres to ethical values. How Does...

---


Across the industry, organizations are attempting to find ways to implement generative AI in their business and operations. But doing so requires significant engineering, quality data and overcoming risks. In this blog post, we show all the elements and practices you need to to take to productize LLMs and generative AI. You can watch the full talk this blog post is based on, which took place at ODSC West 2023, here. Definitions: Foundation Models, Gen AI, and LLMs Before diving into the practice of productizing LLMs, let’s review the basic definitions of GenAI elements: Foundation Models (FMs) - Large deep learning models that are pre-trained with attention mechanisms on massive datasets. These models are adaptable to a wide variety of downstream tasks, including content generation. GenAI / Generative AI - The methods used to generate content with algorithms. Typically, foundation models are used. Large Language Models (LLMs) - A type of foundation model that can perform a variety of NLP tasks. These include generating and classifying text, answering questions in a conversational manner and translating text from one language to another The Landscape: The Race to Open and Commercial LLMs As mentioned, gen AI is garnering industry-wide interest. This has encouraged rapid evolution and intense competition between organizations. New and innovative solutions are constantly emerging, making it hard to pinpoint a single "winning" solution. One notable trend in this space is the rise of specialized or application-specific LLMs. Unlike generalized models, these models are becoming more vertical-oriented, catering to specific...

---


Sense is a talent engagement company whose platform improves the recruitment processes with automation, AI and personalization. Since AI is a central pillar of their value offering, Sense has invested heavily in a robust engineering organization including a large number of data and AI professionals. This includes a data team, an analytics team, DevOps, AI/ML, and a data science team. The AI/Ml team is made up of ML engineers, data scientists and backend product engineers. The Challenge Like many organizations, the AI/ML team at Sense was finding it challenging to scale its ML operations. This was mainly due to three factors: Complexity when managing multiple projects and experiments - Sense had to determine the best strategy for conducting and controlling all their projects and versions, at scale. The need for speed while supporting developer efficiency - Sense needed to ensure fast time-to-market while managing resources efficiently. Establishing a deployment and monitoring strategy - Sense needed to create a sound deployment and monitoring strategy in a cost-effective and straightforward manner. The Solution Sense chose Iguazio as their MLOps solution. Iguazio is an essential component in Sense’s MLOps and DataOps architecture, acting as the ML training and serving component of the pipeline. With Iguazio, Sense’s ML team members can pull data, analyze it, train and run experiments, making the process automated, scalable and cost-effective. “With Iguazio, data scientists and ML engineers start having superpowers. ” Gennaro Frazzingaro, Head of AI/ML at Sense. Iguazio: A Key Component in Sense’s AI/ML and Data Stack...

---


2023 was the year of generative AI, with applications like ChatGPT, Bard and others becoming so mainstream we almost forgot what it was like to live in a world without them. Yet despite its seemingly revolutionary capabilities, it's important to remember that Generative AI is an extension of “traditional AI”, which in itself is a step in the digital transformation revolution. This means that in 2024, we’re likely to see businesses continue to seek ways to adopt generative AI as a way to enhance their operations. But this year, businesses will go beyond the hype. They will focus their resources on optimizing and adapting generative AI, and other AI technologies, attempting to turn them into a driving force for the business. For data science practitioners, productization is key, just like any other AI or ML technology. Successful demos alone just won’t cut it, and they will need to take implementation efforts into consideration from the get-go, and not just as an afterthought. Considerations such as reducing and controlling potential risks, cost effectiveness, scalability, modularity and extensibility, and continuous operations must be part of any implementation.   What will be different in the way businesses approach generative AI this year? What are their expectations from this hyped technology? Will generative AI continue to be one of the hottest topics in 2024 as well? Here are my predictions for the upcoming year: 1. Beyond the Demo: From Prototyping to Generating Business Value The excitement and attention surrounding generative AI is well-deserved, considering...

---


Successfully training AI and ML models relies not only on large quantities of data, but also on the quality of their annotations. Data annotation accuracy directly impacts the accuracy of a model and the reliability of its predictions. This is where human-annotated datasets come into play. Human-annotated datasets offer a level of precision, nuance, and contextual understanding that automated methods struggle to match. In this blog post, we bring the top 16 free human-annotated datasets you can use for your model training and evaluation. To cater to a wide variety of needs, these free datasets for machine learning cover a diverse set of categories and use cases: Sentiment analysis Language and docs For ethical AI use Images and videos When training and developing your models, don’t neglect the final phase - a deployed LLM (or other type of model). Use MLOps solutions to ensure the process is automated, streamlined, scalable and iterative. This will ensure the successful implementation of your model. To learn more about how to build and scale your pipelines, click here. What are Human-annotated Datasets? Human-annotated datasets are data records that have been annotated by humans. This means that humans have added information, like labels or tags. For example, humans can provide inputs for categorization, sentiment analysis, bounding boxes for images, etc. Human annotation helps advance ML and AI model training and evaluation. By providing the ground truth for models, algorithms can understand patterns and make better predictions on new, unseen data. As such, human annotation is...

---


Introducing The New O'Reilly Book: Implementing MLOps in the Enterprise “Implementing MLOps in the Enterprise: A Production-First Approach” is a practical guide, authored by MLOps veterans Yaron Haviv and Noah Gift and published by O’Reilly, which guides leaders of data science, MLOps, ML engineering and data engineering on how to bring data science to life for a variety of real-world MLOps scenarios, including for generative AI. Drawing from their extensive experience in the field, the authors share their strategies, methodologies, tools and best practices for designing and building a continuous, automated and scalable ML pipeline that delivers business value. With practical code examples and specific tool recommendations, the book empowers readers to implement the concepts effectively. After reading the book, ML practitioners and leaders will know how to deploy their ML models to production and scale their AI initiatives, while overcoming the challenges many other businesses are facing. Who This Book Is For This book is for practitioners in charge of building, managing, maintaining, and operationalizing the ML process end to end: Data science / AI / ML leaders: Heads of Data Science, VPs of Advanced Analytics, AI Lead etc. Data scientists Data engineers MLOps engineers / Machine learning engineers This book can also be valuable for technology leaders who want to efficiently scale the use of ML and generative AI across their organization, create AI applications for multiple business use cases, and bridge organizational and technological silos that prevent them from doing so today: CIOs CTOs CDOs Finally, this...

---


An MLOps platform enables streamlining and automating the entire ML lifecycle, from model development and training to deployment and monitoring. This helps enhance collaboration between data scientists and developers, bridge technological silos, and ensure efficiency when building and deploying ML models, which brings more ML models to production faster. When it comes to scaling your MLOps operations, a high-quality, reliable and effective MLOps platform is essential for growth. Some organizations might opt to build one themselves, while others will buy a commercial solution and yet a third group will take a hybrid approach that combines both building and buying. In this blog post, we explore what is required from you for each option and provide tools that will help you make the right choice for your organization to scale your ML and AI activities. Scaling MLOps with Custom Infrastructure Building an MLOps platform might be the right choice for your organization’s growth plans in certain scenarios. Here are some considerations to take into account when you’re contemplating whether to build such a platform. Specialized Requirements - If your ML workflows have unique requirements that off-the-shelf solutions can't address, building your own platform will enable you to customize the solution to your specific needs. Complex Integrations with Existing Systems - If you have a complex ecosystem of tools and databases, building your own MLOps platform may make it easier to achieve seamless integration, as opposed to adapting a commercial solution and getting it to fit into your existing architecture. Data Privacy...

---


Building a smart call center app based on generative AI is a promising solution for improving the customer experience and call center efficiency. But developing this app requires overcoming challenges like scalability, costs and audio quality. By building and orchestrating an ML pipeline with MLRun, which includes steps like transcription, masking PII and analysis, data science teams can use LLMs to analyze audio calls from their call centers. In this blog post, we explain how. The Benefits of Generative AI for Customer Support The number of calls to call centers is increasing, creating a need for improving call center efficiency. AI solutions can help improve operational efficiency by automating processes, streamlining workflows and providing real-time insights. This can help organizations reduce costs, improve customer service and make better decisions. For example, AI solutions can reduce customer frustrations by documenting call details, providing contextual recommendations and analyzing the sentiment of calls. This data can then be leveraged by downstream applications that improve the customer experience, like live agent support, customer profiles, auto-generated content, tailored recommendations, customized offers and more. Challenges of Developing a Call Center Generative AI App Developing a generative AI call center application can improve call center performance, but development comes with its own set of challenges. These include: LLMs for Non-English Languages - LLM performance for non-English languages is usually considered poor. If the call center accepts calls in non-English languages, there needs to be a technological solution to address the gap. Audio Quality - The audio quality...

---


Machine Learning is revolutionizing the world of healthcare. ML models can help predict patient deterioration, optimize logistics, assist with real-time surgery and even determine drug dosage. As a result, medical personnel are able to work more efficiently, serve patients better and provide higher quality healthcare. When developing and training machine learning models for healthcare, open and free datasets are an essential starting point for data scientists and engineers, and they can be hard to come by. Here are 22 excellent open datasets for healthcare machine learning: General Healthcare, Medical and Life Sciences Datasets 1. WHO Global Health Observatory (GHO) resources by the WHO (World Health Organization). The GHO includes data sets and reports from 194 countries on a wide variety of topics. Health topics include mortality, child nutrition, water and sanitation, HIV/AIDS, health systems, injuries, and more. 2. DHS Program  Medical datasets from the DHS (Democratic and Health Services) Program spanning multiple topics. These datasets include data from around the globe, both from individual countries as well as cross-country comparisons. They are based on surveys, biomarker testing and geographic data. 3. HealthData. gov The official US government healthcare website, which includes multiple datasets of the US population. Dataset topics range from COVID-19 to health equity, and more. 4. Life Science Database Archive A life science dataset from Japan, gathered by life scientists over long periods of time. Includes datasets about organs, antigens, chemicals and more. 5. Data. gov. au The official source of Australian open government data. Includes all Australian...

---


ODSC West 2023, one of the leading AI conferences, will take place this year from Oct. 30 to Nov. 2, in San Francisco and virtually. This time, 250 speakers will be sharing 300 hours of valuable content. Sessions will be spread across multiple tracks: NLP and LLMs, MLOps, Generative AI, Machine Learning, Responsible AI, and more. Don’t miss this opportunity to hear from top speakers, get hands-on training, see demos from the latest innovators in the field and network with the data science community. 300 hours of content is a lot, which makes it tough to choose the sessions you’d like to attend. To help, we put together a list of our top 10 recommended sessions. We chose them because they provide practical information that can be implemented today in your environments and organizations, along with a forward-thinking approach that discusses ideas, considerations and opportunities for the future. As can be expected, LLMs and Generative AI are attracting a lot of attention this year, and our list includes sessions about those topics as well. Here are our top recommended 10 sessions: 1. Causality and LLMs Wed. , Nov. 1, 11:20am - 11:50am Robert Osazuwa Ness, PhD Senior Researcher, Microsoft Following the rapid development of LLMs, innovative opportunities for causal analysis have emerged. In this workshop, participants will learn how to effectively use LLMs in complex causal models and in the field of causal AI. First, it will explain how to extract and interpret causal knowledge from LLMs. Then, it will...

---


NLP is a field of AI that enables machines to understand, interpret, and generate human language in a way that is both meaningful and contextually relevant. Recently, ChatGPT and similar applications have created a surge in consumer and business interest in NLP. Now, many organizations are trying to incorporate NLP into their offerings. To help with these efforts, we’ve compiled a list of the top NLP datasets for NLP projects that data scientists and data professionals can use for training their models. This list is a starting point for training your NLP models. The list is divided into a number of groups and types: Q&A Reviews and Ratings Sentiment Analysis Synonyms Emails Long-form Content Audio You can use these datasets for a number of use cases, like creating personal assistants, automating customer service, language translation, and more. The sky's the limit! When planning how to train your NLP models with NLP training datasets, it's important to start with the end in mind — with deployment. To learn more about how to build and scale your NLP pipelines, click here. Now let’s dive into the list: Q&A 1. Stanford Question Answering Dataset (SQuAD) A reading comprehension dataset, comprising pairs of questions and answers based on Wikipedia articles. Get the dataset here. 2. Jeopardy Questions A JSON file with 216,930 Jeopardy questions, answers and additional data like the air date. Get the dataset here. 3. The WikiQA Corpus Question and answer pairs that link to Wikipedia pages with the answer. The data...

---


Generative AI has recently emerged as a groundbreaking technology and businesses have been quick to respond. Recognizing its potential to drive innovation, deliver significant ROI and add economic value, business adoption is rapid and widespread. They are not wrong. A research report by Quantum Black, AI by McKinsey, titled "The Economic Potential of Generative AI”, estimates that generative AI could unlock up to $4. 4 trillion in annual global productivity. However, GenAI usage also needs to be fair, compliant and ethical. One of the main safety concerns is about accidental leaking of PII (Personal Identifiable Information), compromising individuals’ privacy. In this post, we share the open source solution that can help identify and mask PII information: The PII Recognizer. The Challenge: Masking PII PII (Personally Identifiable Information) is any information that can be used to identify an individual. This could include their name, address, phone number, social security number, or credit card number. One of the challenges businesses face when implementing GenAI and training LLMs is accidental exfiltration of PII (Personal Identifiable Information).   LLMs are trained on large datasets of text and code. If this data contains PII, it becomes part of the models’ training dataset. This means it is incorporated into the models and can be used when generating future responses to any user. This could result in a data breach, the compromising of individual privacy and breaking compliance regulations, among other implications. There is also the risk of data retention. Even if the data isn't immediately shared...

---


AI has fundamentally changed the way business functions. Adoption of AI has more than doubled in the past five years, with enterprises engaging in increasingly advanced practices to scale and accelerate AI applications to production. As ML models become increasingly complex and integral to critical decision-making processes, ensuring their optimal performance and reliability has become a paramount concern for technology leaders. This is where model observability and ML monitoring step in, playing a pivotal role in empowering organizations to gain comprehensive visibility into their ML models' behavior and performance in real-world applications. In this article, we delve into the fundamental distinctions between model observability and ML monitoring, shedding light on their unique attributes and functionalities. Moreover, we explore the best practices that enable data scientists and ML engineers to harness the full potential of these practices, ensuring seamless model deployment, rapid issue detection, and continual enhancement. What is Model Monitoring? Model monitoring is a fundamental practice in machine learning that focuses on the systematic observation and evaluation of ML models during their deployment and operation in live applications. As ML models are employed in critical decision-making processes across various domains, ensuring their continued effectiveness and reliability is extremely important. Model monitoring involves the continuous collection and analysis of key performance metrics related to the ML model's behavior, accuracy, and overall performance. These metrics may include prediction accuracy, data and concept drift, feature importance, and model response time, among others. By monitoring these metrics in real-time, data scientists and ML engineers...

---


MLOps accelerates the ML model deployment process to make it more efficient and scalable. This is done through automation and additional techniques that help streamline the process. Looking to improve your MLOps knowledge and processes? You’ve come to the right place. In this blog post, we detail the steps you need to take to build and run a successful MLOps pipeline. What is MLOps? MLOps (Machine Learning Operations) is the set of practices and techniques used to efficiently and automatically develop, test, deploy, and maintain ML models and applications and data in production. An extension of DevOps, MLOps streamlines and monitors ML workflows. With MLOps, organizations can ensure ML projects are delivered efficiently and consistently, while overcoming challenges related to deploying, scaling and maintaining ML models, as well as silos between teams in the organization. As a result, with MLOps - ML models can be deployed much faster to production so they can bring business value. MLOps pipelines support a production-first approach. This means that they are designed to automatically deploy models to production at scale, starting from data collection and all the way to model monitoring. The Challenges MLOps Solves MLOps processes ensure ML models are brought to production and can bring business value. Implementing MLOps solves the following challenges: Siloed Teams - Before MLOps, data scientists, data engineers and DevOps used to work in silos and with different tools and frameworks. Consequently, the models needed to be technologically converted across the different stages, from the lab to production....

---


Generative AI has already had a massive impact on business and society, igniting innovation while delivering ROI and real economic value. According to research by QuantumBlack, AI by McKinsey, titled “The economic potential of generative AI”, generative AI use cases have the potential to add $2. 6T to $4. 4T annually to the global economy. This potential spans more than 60 use cases across all industries. 75% of this impact is focused on marketing and sales, software engineering, customer operations and product and R&D. This is because generative AI has the potential to automate work activities that absorb 60% - 70% of employees’ time today, and especially mundane activities. In this post, we’ll dive into the potential of generative AI, see how organizations should approach leveraging Large Language Models (LLMs) in live business applications. We’ll also discuss how to do it responsibly by embedding Responsible AI principles into generative AI while taking a human-centered approach. This article is based on our recent webinar “MLOps for Generative AI” with guests Nayur Khan, partner at QuantumBlack, AI by McKinsey, Mara Pometti, associate design director, McKinsey & Company, and Yaron Haviv, CTO and co-founder of Iguazio. You can watch the webinar here. The Power and Promise of Generative AI In less than nine months, generative AI has dominated the tech landscape. Tools like Midjourney, Stable Diffusion, Dall-E 2, ChatGPT, Bard and the AI-powered version of Bing have burst onto the scene, revolutionizing our technological capabilities as consumers and business users. They have also...

---


Evaluating ML model performance is essential for ensuring the reliability, quality, accuracy and effectiveness of your ML models. In this blog post, we dive into all aspects of ML model performance: which metrics to use to measure performance, best practices that can help and where MLOps fits in. Why Evaluate Model Performance? ML model evaluation is an essential part of the MLOps pipeline. By evaluating models, data professionals can assess the reliability, accuracy, quality and effectiveness of ML models and ensure they meet the desired technological and business objectives and requirements. In other words, this means ensuring the model answers the business use case as expected and should continue to be deployed in production. In some cases, model performance evaluation can also help meet compliance standards. ML Model Performance Metrics Different metrics can be used to evaluate the performance of ML models. They vary according to the model type and use cases. Therefore, when monitoring the performance of your ML models, choose the relevant ones for your needs. Some of the most common performance metrics for machine learning models include: Classification Model Metrics A classification model is a model that is trained to assign class labels to input data based on certain patterns or features. Common metrics include: Accuracy - The ratio of correctly classified instances to the total number of instances in the dataset. Precision - The proportion of true positive predictions (correctly predicted positive instances) to the total number of positive predictions. Recall (Sensitivity or True Positive Rate)...

---


AutoML (Automated Machine Learning) helps organizations deploy Machine Learning (ML) models faster, by making the ML pipeline process more efficient and less error-prone. If you’re getting started with AutoML, this article will take you through the first steps you need to find a tool and get started. If you’re at an advanced stage, it will help you validate you’re on the right track. After reading this article, you’ll have a good understanding of the benefits of AutoML, the available tools and how to choose the right tool for your industry. What is Automated Machine Learning? Automated Machine Learning (AutoML) is the automation of part of or all of the ML pipeline, in order to increase productivity and accuracy and eliminate organizational silos. ML pipeline stages that can be automated include model selection, the selection and parametrization of the ML algorithm for model training, data preprocessing, feature engineering, hyperparameter tuning and deployment. Tasks that can be automated include the selection of appropriate algorithms and feature sets, scaling numerical features, handling missing values, encoding categorical variables, testing and hypertuning parameters. By automating the complex and time-consuming tasks and activities required for building and deploying ML models, machine learning automation creates standardization. This makes these expert-created processes accessible to a wider range of users and increases efficiency. As a result, AutoML accelerates the development and deployment of ML models and makes them useful for business use cases. What are the Advantages of Automated Machine Learning Tools? AutoML tools have become increasingly popular in...

---


How to train on Databricks and deploy with MLRun Every organization aiming to bring AI to the center of their business and processes strives to shorten machine learning development cycles. Even data science teams with robust MLOps practices struggle with an ecosystem that is in a constant state of change and infrastructure that is itself evolving. Of course, no single MLOps stack works for every use case or team, and the scope of individual tools and platforms vary greatly. Databricks is a hugely popular data platform with core ML capabilities, reportedly used by over 5000 enterprises. But for use cases requiring automation and scale, or complex use cases with real-time applications, data science teams may need some extra tooling for model serving. To understand why you might want that extra tooling, consider a complex but fairly common scenario: fetching historical data for the training side, with the security requirement that live data remains on-prem. Databricks does offer real-time model serving capabilities, but by using MLRun real-time serving with the complex serving graph, you can deploy the real-time serving with pre-processing, model inference, post-processing in the same graph and scale different part of the graph independently. In this blog post, I’ll show how to integrate models trained in Databricks and perform model serving with open-source tool MLRun. What is MLRun and the Iguazio MLOps Platform? MLRun is a popular open-source MLOps framework that streamlines machine learning projects from data collection, experimentation, model development, feature creation, production model serving deployment and model...

---


Deploying trained models takes models from the lab to live environments and ensures they meet business requirements and drive value. Model deployment can bring great value to organizations, but it is not a simple process, as it involves many phases, stakeholders and different technologies. In this article, we provide recommendations for data professionals who want to improve and streamline their model deployment process. This list is based on our experience deploying models for large and small organizations across the globe.   Overview of Deploying a Machine Learning Model Deployment processes will differ between organizations, depending on their technologies, environments and use cases. However, in most organizations, there are a few steps that will always repeat themselves. These steps include: Packaging: Adding dependencies and parameters, running scripts and performing builds. Scaling: Managing load balancing, data partitions, model distribution and AutoML. Tuning: Performing data parallelism, managing GPUs, tuning queries and caching. Instrumentation: Monitoring, logging, versioning and managing security. Automation: Of CI/CD, workflows, rolling upgrades and A/B testing. To minimize friction and errors and increase the chance of deployment success, it is recommended to streamline the machine learning deployment process by automating it as much as possible. An efficient and comprehensive model deployment framework will go a long way to support this. Data Preparation One of the most important steps that support successful model deployment is data preparation. In this stage, the data is pre-processed so it can be used in the deployed model. Proper data preparation ensures the model provides accurate and...

---


ODSC East is one of the leading data science conferences. This year, it’s taking place in Boston and virtually, between May 9 and 11. The event will be packed with valuable content, with 250 speakers across tracks like Machine Learning, Hands-on Training, Deep Learning, NLP, Responsible AI, and many more. Don’t miss this opportunity to stay up-to-date on the latest technological advancements, build your job skills, and network peers. With so many options, identifying the sessions you’d like to attend is challenging. To help, we gathered this list of our top seven recommended sessions. We chose them based on their ability to provide practical, yet innovative, ways to solve challenges encountered by data professionals. Here are our top recommended 8 sessions: 1. The Data Cards Playbook: A Toolkit for Transparency in Dataset Documentation Wed. , May 10, 2:20pm - 3:50pm Andrew Zalvidar, Senior Developer Relations Engineer, Google Research Data Cards are transparency artifacts with the structured summaries of ML datasets, explanations of the processes and logic that shape the data and descriptions of how the data can be used to train and evaluate ML models. In this session, Andrew Zalvidar from Google Research shares the Data Cards Playbook, a toolkit that provides activities, frameworks and guidance designed to help teams and organizations in their dataset transparency efforts. This workshop will guide participants about how to use this open-source and self-service kit. It will also provide evidence-based patterns to help predict and avoid common challenges. Read the full abstract here. 2....

---


The open source ML tooling ecosystem has become vast in the last few years, with many tools covering different aspects of the complex and expansive process of building, deploying and managing AI in production. Some tools overlap in their capabilities while others complement each other nicely. In part because AI/ML is still an emerging and ever-evolving practice, the messaging around what all these tools can accomplish can be quite vague. In this article, we’ll dive into three tools to better understand their capabilities, the differences between them, and how they fit into the ML lifecycle. Kubeflow, MLflow, and MLRun are popular open source tools in the ML ecosystem with similar attributes, that actually address different facets of the ML lifecycle. How are they alike? Each of them enables cross-functional collaboration, with some level of parameter, artifact, and model tracking. Each is supported and maintained by major players in the AI industry. Each of them addresses data science and experimentation requirements While there are a handful of feature-level similarities among these three tools, it’s important to understand that they each solve different challenges, and one is not a true replacement for the other. What is Kubeflow? Kubeflow, started as an internal tool at Google and is now a multi-architecture, multi-cloud framework, which is described as the ML toolkit for Kubernetes. It provides a system for managing ML components on top of Kubernetes and acts as the bricks and mortar for model development, with a focus on the automation, scaling and tracking...

---


Transforming Manufacturing at Seagate: Using AI to Detect Defects on the Factory Floor Seagate wanted to leverage their vast petabytes of sensor and image data to reduce chip manufacturing times and capital costs and maintain quality, but faced challenges: Efficient data processing methods (compute and orchestration) while handling petabytes of real-time data Long manufacturing cycles, with 40% of time spent on measurement Sampling, logistics and time limitations With Iguazio, they were able to transition from a manual inspection process via microscope to fully automated deep learning and computer vision inspection. By leveraging fully managed data engineering and automation, they were able to: Doubled first-pass wafer yield and eliminated contamination by automating measurement and implementing deep learning visual inspection steps Reduced manufacturing steps by 40%, thereby reducing capital costs by modularizing and streamlining data engineering processes. Improved cloud usage efficiency by 6x. Business Background Seagate is the world’s leading data storage solution. Together with Iguazio, Seagate is able to manage data engineering at scale while harnessing petabytes of data, efficiently utilize resources, bridge the gap between data engineering and data science and create one production-ready environment with enterprise capabilities. In this new webinar, Vamsi Paladugu, Sr. Director of Lyve Cloud Analytics at Seagate explains about Seagate’s partnership with Iguazio, shares how Iguazio helped them overcome business and technical challenges and dives into a real-life use case where Iguazio helped leverage ML to improve productivity and cut costs. Watch the webinar here. About Seagate Seagate is the data storage industry leader, powering...

---


8 years ago, when I founded Iguazio together with my co-founders Yaron Haviv, Yaron Segev & Orit Nissan-Messing, I never thought I would be making this announcement on our company blog: McKinsey acquired Iguazio! When we first embarked on this journey, we realized that while AI has the ability to transform any industry - from banking to retail to manufacturing - in reality most data science projects fail. In fact, according to McKinsey research, over $490B have been invested in AI between the years 2012-2021, with little return. Only one out of ten attempts to deploy AI in enterprise-level operations succeeds - 90% of AI projects never even make it out of the lab. This gap between the great potential of AI and its actual business impact was our north star as we built and expanded the Iguazio MLOps Platform, adding features and functionality to automate, accelerate and simplify the data science process. We set on a mission to change this reality for enterprises, and I’m happy to say we’ve been fortunate enough to work with some incredible clients and see some great success stories with results such as a 12X acceleration of the data science process, 50% reduction in cost and 90% reduction in code. Managing a startup is a roller coaster – The highest highs and the lowest lows. From tremendous technological achievements, incredible partnerships, industry alliances and events with standing-room-only - to a global pandemic, economic recession, and everything in between. It’s been a heck of a ride....

---


HCI (Home Credit International) is a global consumer financial provider. As leaders in their space, they identified the potential of ML models in financial institutions, and especially in risk-related use cases. However, deploying ML models in enterprises is not always an efficient process: time to delivery is long and access to data is limited. Jiri Steueur, Enterprise Architect at HCI, recently joined us for a webinar to share his top tips and ideas for achieving MLOps efficiency. In this blog post, we bring a concise overview of his findings. To see Steuer’s entire presentation, including an in-depth description of an MLOps efficiency solution, you can watch the entire webinar here. Machine Learning Use Cases in Financial Institutions There is high potential for ML in financial institutions, especially with regards to risk-related requirements. Some of the top use cases include: Campaign management - Identifying the most fitting consumers for products Generating offers - Defining default and pre-approval limits Risk-based pricing - Calculating prices while balancing attractive products with product security Next best offer (cross selling) with the prediction of the next best offer Next best offer with the ability to decrease insurance-based risks Penalties when servicing products Payment behavior, calendars and promise to pay during collection processes Anti-fraud to protect clients, through device fingerprinting, mobile device scanning, malware detection, and more Creating a client behavioral profile based on predictions from data mining Improving ML Efficiency Yet, despite the value of ML, and according to HCI’s internal research, nearly 80% of the...

---


Enterprises who are actively increasing their AI maturity in a bid to achieve business transformations often find that with increased maturity comes increased complexity. For use cases that require very large datasets, the tech stacks required to meet business needs quickly become unwieldy. AI services that involved predictive maintenance, fraud detection, and NLP serve critical business functions, but under the hood, the data wrangling required is a major bottleneck for the teams tasked with delivering business value. Distributed ingestion is one piece of the data complexity puzzle. With distributed ingestion tools like Spark, data science teams can increase scalability to process large amounts of data quickly and efficiently, improve performance by speeding up the overall time it takes to ingest and process the data, and leverage a simple API to abstract away much of the complexity of using multiple machines. But distributed ingestion tool is just one component of an end–to–end ML pipeline, so it needs to work together with all the other parts. A tech stack for a particularly complex use case could include components like a feature store, a data lake and much more. Adding a tool to the stack is never simple--integrating and maintaining multiple services can become an overwhelming task for engineering teams.   For enterprises who are deploying AI at scale, a feature store has become a critical component. The Iguazio feature store is integrated with popular tools such as Spark, Dask, Snowflake, BigQuery, and more. In addition to integration with external tools, Iguazio’s feature...

---


In 2022, AI and ML came into the mainstream consciousness, with generative AI applications like Dall-E and GPT AI becoming massively popular among the general public, and ethical questions of AI usage stirring up impassioned public debate. No longer a side project for forward-thinking businesses or CEOs that find it intriguing, AI and ML are now moving towards the center of the business. For enterprises, this means that in 2023 AI and ML have the potential to exceed every technological and business expectation. As an ML industry veteran, I’m excited to see the rapid changes in AI and ML’s technological capabilities and the accelerated adoption rate in the past years. As we raise our glasses to the upcoming year, here are my predictions of what we’re expected to face as an industry in 2023: From ML Models to AI Apps Up to now, the popular approach has been to start with building models and thinking about the overall application and business integration later. Models were used initially by business analysts and reporting systems which didn’t require integration with business applications. As AI becomes central to the business, data science teams now need to integrate with live data sources and existing applications, and turn manually executed applications into interactive and real time applications. The traditional approach is to produce model serving endpoints. These endpoints accept a numeric feature vector and respond with a prediction output. In this scenario, the critical functionality of integrating with data, adding business logic, acting on the...

---


The IDC MarketScape: Worldwide Machine Learning Operations Platforms 2022 Vendor Assessment is an annual study that evaluates technology vendors based on a comprehensive framework. It provides an in-depth quantitative and qualitative assessment of MLOps solution vendors in a long-form research report, to help buyers make important technology decisions that will create long term business success. We're proud to announce that Iguazio has been named a Major Player in the IDC MarketScape: Worldwide Machine Learning Operations Platforms 2022 Vendor Assessment. This report is the latest recognition from several top industry analyst firms for the Iguazio MLOps Platform. The Iguazio MLOps Platform was Recognized for Enterprise AI Acceleration As AI adoption increases among enterprises, the need to operationalize ML presents a number of obstacles, including legacy tech stacks inadequate for running ML at scale, a lack of collaboration between different departments, ad-hoc processes and insufficient automation. Enterprises that are serious about AI are now looking to build ‘ML factories’—where repeatable and reliable workflows can help quickly roll out new AI services to production.   To build such an 'ML factory', data science teams are looking for solutions to automate and accelerate machine learning projects, deploy them to production, monitor and manage them in live business environments. The Iguazio MLOps Platform enables multi-functional data science teams to develop, deploy and manage AI applications effectively and efficiently, even at scale and in real-time environments. Iguazio enables data science, data engineering and ML engineering teams to work on a single platform to operationalize machine learning...

---


The GigaOm Radar reports support leaders looking to evaluate technologies with an eye towards the future. In this year's Radar for MLOps report, GigaOm gave Iguazio top scores on multiple evaluation metrics, including Advanced Monitoring, Autoscaling & Retraining, CI/CD, and Deployment. Iguazio was therefore named a leader and also classified as an Outperformer for its rapid pace of innovation. The goal of MLOps is to make ML a fully integrated part of business operations. Enterprise MLOps involves implementing machine learning applications in live business environments, to solve specific business problems, often across multiple teams. This requires a platform with enterprise-grade capabilities, not only for development, but also for deployment at scale and post-deployment monitoring and management of models in production, in live business environments. The GigaOm Radar for MLOps helps buyers become familiar with the growing range of MLOps solutions and vendor offerings. Why GigaOm Named Iguazio an Outperforming Leader for 2022 Iguazio is honored to be named an Outperforming Leader in GigaOm’s latest MLOps report. This recognition highlights our rigorous production-first approach to MLOps and differentiated capabilities that address the entire AI/ML lifecycle. We're proud to see that the Iguazio MLOps Platform has been recognized for its ability to help enterprises scale, automate and accelerate enterprise AI, with features such as: CI/CD for ML Real-time serving pipelines An integrated online and offline feature store Built-in monitoring and retraining What the Iguazio MLOps Platform Brings to Enterprise AI As organizations continue to adopt machine learning, technical teams are shifting their...

---


Hugging Face is a popular model repository that provides simplified tools for building, training and deploying ML models. The growing adoption of Hugging Face usage among data professionals, alongside the increasing global need to become more efficient and sustainable when developing and deploying ML models, make Hugging Face an important technology and platform to learn and master.   Together with MLRun, an open source platform for simplifying the deployment and management of MLOps pipelines, Hugging Face enables data scientists and engineers to get their models to production faster and in a more efficient manner. In this blog post, we introduce Hugging Face and MLRun and show the value of running them together. This blog is based on the webinar “How to Easily Deploy Your Hugging Face Models to Production”. The webinar also shows a live demo of a Hugging Face deployment with MLRun, including data preparation, a real application pipeline, post-processing and model retraining. You can watch the webinar, presented by Julien Simon, Chief Evangelist at Hugging Face, Noah Gift, MLOps Expert and author, and Yaron Haviv, co-founder and CTO of Iguazio, here. How Transformers Have Revolutionized Deep Learning One of the most recent trends in ML is the reinvention of deep learning. Or rather, it is more accurate to say that transformer models are “swallowing up” traditional deep learning. Traditional deep learning architectures like CNNs, LSTMs and RNNs, which attempt to solve problems with unstructured data like audio and images, were very popular until recently.   But while these...

---


ODSC West is a fascinating three-day professional event, taking place in San Francisco and virtually, from Nov. 1 - 3, 2022. Topics like ML and deep learning, data engineering and MLOps, NLP and Responsible AI will be discussed and showcased during ODSC West through training sessions, workshops, bootcamps and keynote talks. As part of our pre-event tradition here at Iguazio, we’re excited to share with the ML and AI community our list of recommended sessions. We chose them based on their novel approaches to ML and AI, their diverse ways of approaching data science and their applicability to a wide variety of business and real-life use cases. Here are our top recommended 10 sessions: 1. Hyper-productive NLP with Hugging Face Transformers Tue. , Nov. 1, 9:40am - 10:55am A workshop showing how to use Hugging Face Transformers, the popular open source NLP library. During the workshop, Julien Simon, Chief Evangelist at Hugging Face, will show how to train and deploy a text classification model that predicts ratings based on a data set with Amazon product reviews. The code from the session will also be shared with attendees and available for reuse. 2. Self-Supervised and Unsupervised Learning for Conversational AI and NLP Tue. , Nov. 1, 11:25 AM - 12:40 PM Self-supervised and Unsupervised learning techniques are restructuring AI in fields like NLP, Computer Vision and Robotics. In this workshop, Chandra Khatri, Chief Scientist and Head of AI at Got it AI, will show how to leverage transformers and large language...

---


With the Apache Spark 3. 1 release in early 2021, the Spark on Kubernetes project has been production-ready for a few years. Spark on Kubernetes has become the new standard for deploying Spark. In the Iguazio MLOps platform, we built the Spark Operator into the platform to make the deployment of Spark Operator much simpler.   This blog post will cover the benefits of running your workload on Spark Operator, and how to run the workloads on Spark Operator with dynamic allocation of executors with MLRun, Iguazio’s open source MLOps orchestration framework.   Spark Operator Concept Here is a quick introduction to how Spark Operator works. For more details, you can refer to the Spark Documentation  Spark creates a Spark driver running within a Kubernetes pod.   The driver creates executors which are also running within Kubernetes pods and connects to them, and executes application code.   When the application is complete, the executor pods terminate and are cleaned up, but the driver pod persists logs and remains in “completed” state in the Kubernetes API until it is eventually garbage collected or manually cleaned up.   Note that in the completed state, the driver pod does not use any computational or memory resources. The driver and executor pod scheduling are handled by Kubernetes. Communication to the Kubernetes API is done via fabric8. It is possible to schedule the driver and executor pods on a subset of available nodes through a node selector using the configuration property for it. It will be...

---


When operationalizing machine and deep learning, a production-first approach is essential for moving from research and development to scalable production pipelines in a much faster and more effective manner. Without the need to refactor code, add glue logic and spend significant efforts on data and ML engineering, more models will make it to production and with less issues like drift. In this blog post, we’ll demonstrate how to use Iguazio & Snowflake to create a simple, seamless, and automated path to production at scale. This post is based on a talk I gave at MLOps NYC Summit. Prefer to see the video? You can watch the whole thing here: hbspt. cta. load(5272961, '8b366c1e-a20a-45ed-820c-2c4f41fde06f', {"useNewLoader":"true","region":"na1"}); The MLOps Pipeline: A Short Intro Before diving into how the Iguazio feature store and Snowflake operate together, let’s review why we need a solution for automating the ML pipeline in the first place. Deploying machine learning models to production introduces four main challenges: 1. Siloed Work - The machine learning pipeline requires collaboration between data scientists, data engineers and DevOps. Today, these teams often work in silos and use different tools and frameworks. As a result, code and models that were developed in the lab need to be technologically converted so they match the production environment’s requirements. This process creates friction that results in some models not making it to production. 2. Lengthy Processes - The route from the lab to production is long and time-consuming. It constitutes multiple phases, including testing, security, versioning, tuning,...

---


Iguazio users can now run their ML workloads on AWS EC2 Spot instances. When running ML functions, you might want to control whether to run on Spot nodes or On-Demand compute instances. When deploying Iguazio MLOps platform on AWS, running a job (e. g. model training) or deploying a serving function users are now able to choose to deploy it on AWS EC2 Spot compute instances. Choosing a spot instance is a great cost saving choice if you can be flexible about when your applications run and if your applications can be interrupted from your ML perspective. Configuring Spot nodes inside the Iguazio MLOps Platform About Amazon EC2 Spot Instances: “Amazon EC2 Spot Instances let you take advantage of unused EC2 capacity in the AWS cloud. Spot Instances are available at up to a 90% discount compared to On-Demand prices. You can use Spot Instances for various stateless, fault-tolerant, or flexible applications such as big data, containerized workloads, CI/CD, web servers, high-performance computing (HPC), and test & development workloads. Because Spot Instances are tightly integrated with AWS services such as Auto Scaling, EMR, ECS, CloudFormation, Data Pipeline and AWS Batch, you can choose how to launch and maintain your applications running on Spot Instances. Moreover, you can easily combine Spot Instances with On-Demand, RIs and Savings Plans Instances to further optimize workload cost with performance. Due to the operating scale of AWS, Spot Instances can offer the scale and cost savings to run hyper-scale workloads. You also have the option to hibernate, stop or...

---


AutoML with experiment tracking enables logging and tracking results and parameters, to optimize machine learning processes. But current AutoML platforms only train models based on provided data. They lack solutions that automate the entire ML pipeline, leaving data scientists and data engineers to deal with manual operationalization efforts. In this post, we provide an open source solution for AutoMLOps, which automates engineering tasks so that your code is automatically ready for production. For more in-depth explanations and a live demo of the solution, you can watch the webinar this blog post is based on, here. Experiment Tracking: Benefits and Challenges Traditionally, machine learning models are developed by data scientists and engineers obtaining data from a data warehouse, a data lake or standalone files. They need to prepare it, train their model or use an AutoML service, evaluate the model for accuracy and finally generate a model for production. This process needs to be tracked, so data professionals can understand how any parameter changes they make impact the model’s behavior and accuracy. Data scientists often use experiment tracking for monitoring and logging model results, since it provides them with better decision making (understanding which parameters and inputs yield the best results and reusing them in future models) and visualization (for better understanding the experiment results and model performance evaluation). Experiment tracking can also help with tracking additional aspects of the ML workload: Accelerating the process of moving results from a successful experiment to productionSimplifying pipelines - building and deploying pipelines in...

---


In this article, we will walk you through steps to run a Jenkins server in docker and deploy the MLRun project using Jenkins pipeline.   Before we dive into the actual set up, let’s have a brief background on the MLRun and Jenkins.   What is MLRun?   MLRun is a popular open source MLOps orchestration framework that streamlines machine learning projects from data collection, experimentation, model development, feature creation, production model serving deployment and model monitoring, and the full lifecycle management of machine learning. Iguazio MLOps Platform is built with MLRun at its core, with added enterprise features such as data management, user management, real-time model serving, security, autoscaling, high availability and more. Please see mlrun and the Iguazio MLOps Platform for more information. What is Jenkins Pipeline?   A continuous delivery (CD) pipeline is an automated expression of your process for getting software from version control right through to your users and customers. Every change to your software (committed in source control) goes through a complex process on its way to being released. This process involves building the software in a reliable and repeatable manner, as well as progressing the built software through multiple stages of testing and deployment.   In a traditional build process, the code and build process are separated, code is in the SCM system while the build process is residing inside the Jenkins configuration in a Jenkins server. This process had a lot of management challenges. As we are moving towards CI/CD “as code”,...

---


We’re so proud to share that Iguazio has been named a sample vendor in eight Gartner Hype Cycles in 2022: The Hype Cycle for Data Science and Machine LearningThe Hype Cycle for Artificial IntelligenceHype Cycle for Analytics and Business IntelligenceThe Hype Cycle for Infrastructure StrategyThe Hype Cycle for ITSMHype Cycle for Healthcare Data, Analytics and AIThe Hype Cycle for Customer Experience AnalyticsThe Hype Cycle for CRM Sales Technology Iguazio was mentioned in the following categories: MLOps, Logical Feature Store, Adaptive ML, Data-Centric AI, AI Engineering, AI TRiSM, Operational AI Systems, ModelOps, AI Engineering in HCLS and Continuous Intelligence. We are delighted to have been mentioned alongside global industry leaders like AWS, IBM, Microsoft, Google, Databricks and Dataiku. These hype cycles are a great tool for enterprises considering their AI methodology and evaluating vendors. Here are some details on each report: 1. The Hype Cycle for Data Science and Machine Learning, 2022 analyzes the maturity of the DSML landscape and how it is evolving to meet the requirements of the enterprise while delivering business value. According to the DSML Hype Cycle, innovation and novel techniques are being used by data and analytics leaders to find solutions to challenges and to stay ahead of the pack. In this DSML Hype Cycle, Iguazio is mentioned in the following categories: MLOps, which discusses the streamlining of the end-to-end development, testing, validation, deployment, operationalization and instantiation of ML models; and how MLOps is standardizing the process to drive ML value. Logical Feature Store, which explains...

---


What is an AI App? Machine learning is more accessible than ever, with datasets available online and Jupyter notebooks providing an easy way to explore and train models with minimal expertise. But when building a model, it's easy to forget that its real value lies in being incorporated into a live application that will provide value to the user. Therefore, we wanted to demonstrate how we can very easily leverage the models we build into a full application, with minimal engineering. In this blog, we'll demonstrate how quickly and easily you can create what we’re calling an AI App in Under 20 minutes. The idea is to create a set of simple AI applications that leverage a machine learning model and allow you to interact with a UI to visualize the results of their predictions. We will show that you don't have to build the whole pipeline from scratch to deploy and interact with these models. Along with interactive UIs, these AI apps will also contain Behind The Scenes  code snippets showing how the models were deployed and how the application interacts with them. How Do We Embed AI into an Application? The tool that makes this all possible is MLRun - an end-to-end, open-source MLOps orchestration framework that includes experiment tracking, job orchestration, model deployment, feature store, and much more. Our focus in this blog will be on MLRun's model deployment capabilities. In machine learning, deployment refers to the process of making a model available for use. As a...

---


Why Experiment Tracking  In your machine learning projects, have you ever wondered “why is model Y is performing better than Z, which dataset was model Y trained on, what are the training parameters I used for model Y, and what are the model performance metrics I used to select model Y? ” Does this sound familiar to you? Have you wondered if there is a simple way to answer the questions above?  Data science experiments can get complex, which is why you need a system to simplify tracking. In this blog, you will learn how you can perform all those tasks (as well as set up an automated machine learning pipeline) with a few simple steps using MLRun.   Automated Experiment Tracking with MLRun  MLRun is an open source framework to orchestrate MLOps from the research stage to production-ready AI applications. With a feature store and a modular strategy, MLRun enables a simple, continuous, and automated way of creating scalable production pipelines. MLRun automates the build process, execution, data movement, scaling, versioning, parameterization, outputs tracking, CI/CD integration, deployment to production, monitoring, and more. MLRun empowers multiple-functional ML teams to collaborate on the orchestration of the entire ML lifecycle and automate MLOps. MLRun simplifies and accelerates the time to production. To learn more about the MLRun open source project, please refer to this git repo.   In this article, you will see how to turn your existing model training code into an MLRun job and get the benefit of all the...

---


We’re proud to share that Iguazio has been named in Gartner's 2022 Market Guide for Data Science & Machine Learning Engineering Platforms. According to Gartner, “The AI & data science platform market is due to grow to over $10 billion by 2025 at a 21. 6% compounded annual growth rate. This growth mirrors the investments made by organizations in DSML initiatives, which are largely turning from strategy to execution. ” This new Gartner guide is intended for technology leaders, to gain insights into the DSML platform market and help them choose the right solution for their specific needs. The guide assesses engineering requirements and provides a framework for vendor comparison and selection. Being one of the first vendors in the MLOps market, with highly differentiated capabilities and an aggressive approach to innovation, Iguazio is uniquely positioned to help enterprises bring data science to production. We’re proud to see the Iguazio MLOps platform recognized by customers, partners and the industry for capabilities such as: End-to-end MLOps automation, CI/CD for ML, real time serving pipelines, a built in feature store, real-time feature engineering, model monitoring and more. Iguazio is used by companies like the LATAM Airlines Group, which is rolling out a large-scale, cross-company AI innovation project using Iguazio on GCP, alongside many other key customers across verticals. The Iguazio MLOps platform abstracts away the complexities of MLOps and empowers enterprises to create real business impact with AI, across a variety of use cases. In addition to Iguazio being mentioned in this new Gartner guide, Iguazio...

---


As a very hands-on VP of Product, I have many, many conversations with enterprise data science teams who are in the process of developing their MLOps practice. Almost every customer I meet is in some stage of developing an ML-based application. Some are just at the beginning of their journey while others are already heavily invested. It’s fascinating to see how data science, a once commonly used buzz word, is becoming a real and practical strategy for almost any company. In the following post, I’ll address one of challenges that customers bring up time and again: running and tuning data science experiments. With a step by step tutorial I’ll cover complexity concerns and show how to solve them with MLRun, our open source MLOps orchestration framework which enables a simple, continuous, and automated way of creating scalable production pipelines. MLRun is all about automation: with a few simple lines of code, it automates the build process, execution, data movement, scaling, versioning, parameterization, outputs tracking, CICD integration, deployment to production, monitoring, and more. MLRun enables data science teams to track machine learning experiments by providing a generic and easy to use mechanism to describe and track code, metadata, inputs and outputs of machine learning related tasks (executions). Machine learning experiment management is part of any healthy MLOps process. MLRun tracks various elements, stores them in a database and presents all running jobs as well as historical jobs in a single report. The database location is configurable and users run queries to...

---


ODSC Europe 2022 will be an exciting hybrid event, taking place in London and virtually, from June 15-16. With 50 workshops, 150 sessions and thousands of attendees, it is surely the place to learn about AI and ML and meet colleagues and experts from around the globe. Choosing which sessions to attend is not always easy, so we put together this list of our top recommended sessions for the event. We chose them based on their innovative way of thinking and the new approaches and methodologies they offer, while remaining connected to real-life examples and challenges. Here are our top recommended 9 sessions, in order of appearance: 1. Keynote: MLOps Beyond Training: The Production-First Approach to AI Wed. , June 15, 10:00am - 10:40am AI journeys usually start by manually building AI models. But to create business value, we need to take a production-first approach to MLOps. This means designing a continuous operational pipeline, automating components and measuring metrics. In this keynote session, Yaron Haviv, Iguazio’s co-founder and CTO, will explain what it means to operationalize ML, the challenges and how to overcome them. The session will include an overview of all the required steps alongside real-world examples. 2. Explainability by Design: a Methodology to Support Explanations in Decision-making Systems Wed. , June 15, 10:50am - 11:30am How can we ensure automated decision-making is meaningful? A new method, explainability by design, introduces proactive measures to include explanations in the design, instead of treating it as an afterthought. In this eye-opening...

---


Data science is an important skill, but the hard truth is many organizations aren’t seeing the ROI showing that data science work is making a business impact. Yet today, many organizations are still struggling to adopt a holistic approach centered around creating business value. Instead, they are focused on theoretical work. Here at Iguazio, we recently held a webinar with Noah Gift, founder of Pragmatic A. I. Labs, professor, author and MLOps consultant. In his talk, he provided best practices for succeeding with MLOps and connecting data science to a clear ROI. Below, you can find the main best practices and future trends he presented. The entire webinar is available on-demand, and has more in-depth explanations and examples. 1. Pick the Right Technology Partners and Solutions for You The right technology partners will help you scale, adapt to future technological changes and be innovative. We recommend having two to three technology partners/solutions in place: The first, your foundational solution, which is meant to be your primary product. These are solutions like AWS, Azure or Google Cloud. The second solution is a platform that solves a specific problem very well. For example, Iguazio, Splunk, Snowflake or Databricks. The third is a long-term R&D investment. For example, frameworks for ML/DL, Kubernetes or edge computing. How should you choose these primary, secondary and investment partners? While there is no silver bullet for choosing the right platform or technology provider, there is a set of considerations Gift recommends taking into account. Considerations for choosing...

---


MLOps World is taking place this year in Canada and virtually, from June 7 to 10, 2022. Packed with more than 80 workshops and case study talks, MLOps World is an important global gathering for anyone working with ML or AI. With so many incredible sessions, it can be hard to choose which ones to attend. To help, we compiled our top recommended ones. We chose them based on the use cases they cover, their comprehensiveness of productizing ML and the practical tools they offer. We will also be there, more details below. Here are our top recommended eight sessions, in the order of their appearance: 1. Machine Learning Monitoring in Production: Lessons learned from 30+ Use Cases June 7th, 10am - 12pm Machine learning monitoring is essential for detecting problems in ML stacks. But how should data scientists and engineers get started? In this session, Lina Weichbrodt from DKB Bank will explain about the four golden signals, how to prioritize a service response and how to monitor with the existing tools you already have! 2. Implementing MLOps Practices on AWS using Amazon SageMaker June 7, 10am - 12:30 pm A hands-on workshop by Shelbee Eigenbrode, Bobby Lindsey and Kirit Thadaka from AWS, teaching how to use Amazon SageMaker Pipelines to implement ML pipelines with a CI/CD approach. 3. Production ML for Mission-Critical Applications June 7, 1pm - 1:40pm How does Google implement production applications in ML pipeline architectures while ensuring they are production-ready? In this talk, Robert Crowe from...

---


Many organizations are turning to Snowflake to store their enterprise data, as the company has expanded its ecosystem of data science and machine learning initiatives. Snowflake offers many connectors and drivers for various frameworks to get data out of their cloud warehouse. For machine learning workloads, the most attractive of these options is the Snowflake Connector for Python. Snowflake added some new additions to the Python API in late 2019 that improved performance when fetching query results from Snowflake using Pandas DataFrames. In their blog post just a few months later, internal tests showed a 10x improvement if you download directly into a Pandas DataFrame using the new Python client APIs. In this article, we’re going to show you how to use this new functionality. Install the Snowflake Connector Python Package pip install snowflake-connector-python The snowflake connector needs some parameters supplied so it can connect to your data. Snowflake user, password, warehouse, and account all need to be supplied. In our code examples you will see these params given as **connection_info for brevity. Below is a very simple example of using the connector. This uses the fetch_pandas_all function, which retrieves all the rows from a SELECT query and returns them in a pandas dataframe. To be clear, this does not replace the pandas read_sql method, as it only supports SELECT statements. import snowflake. connector as snow ctx = snow. connect(**connection_info) query = "SELECT * FROM SNOWFLAKE_SAMPLE_DATA. TPCH_SF1. CUSTOMER" cur = ctx. cursor cur. execute(query) df = cur. fetch_pandas_all This example...

---


On April 19, 2022 data scientists, data engineers and AI professionals will gather in Boston and virtually, to attend ODSC East. Over the course of three days, from April 19 to April 21, expert speakers will discuss deep learning, ML, MLOps, data visualization, responsible AI, and more. This year, many of the talks focus on governance and security, as well as various interesting use cases. Since there are so many great sessions and it’s so hard to choose which ones to attend, we’ve put together our recommendations. We will also be there - more details below. Here are our top 11 recommended sessions for ODSC East: 1. Adversarial Robustness: How to Make Artificial Intelligence Models Attack-proof! Serg Masís, Climate Data Scientist, Syngenta April 19, 2PM - 4PM A security-focused session about evasion attacks, which occur when perpetrators trick classifiers to make false predictions. The session will examine an evasion attack use case and explain two defense methods for ML models. Then there will be a demonstration of a robustness evaluation method and a certification method, to assure the ML model will resist evasion attacks. 2. ODSC Keynote - The Big Wave of AI at Scale Luis Vargas, Partner Technical Advisor, Microsoft April 19, 9 AM - 9:40 AM In this session, the Partner Technical Advisor to the CTO of Microsoft discusses the trend of larger AI models that enable new tasks in language, vision, and multi-modality. He will provide an overview of the research and engineering efforts that comply with...

---


This tutorial demonstrates the availability of streaming data in a data science environment, which is useful for working with real-time and fresh datasets: First, we collect data from an existing Kafka stream into an Iguazio time series table. Next, we visualize the stream with a Grafana dashboard; and finally, we access the data in a Jupyter notebook using Python code. We use a Nuclio serverless function to “listen” to a Kafka stream and then ingest its events into our time series table. Iguazio gets you started with a template for Kafka to time series. We visualize the data with Grafana and work with time series data using Python code in Jupyter. Data scientist easily access both historical and real-time data in a full Python environment for exploration and training with Iguazio. If you'd prefer to read along, here's a transcript of how we did it: Here’s how to make streaming data available for data scientists in just a few minutes. We're going to collect data into an Iguazio time series table from an existing Kafka stream, and then create a quick Grafana dashboard and access the data using Python code from a Jupyter notebook. The idea here is to have data scientists getting real-time and fresh data into their data science environment where they can immediately start working with streaming datasets. Create the Time Series Table Go to the services view and then to the shell service. I'm using tstb CLI. Specify the container name and table name with the...

---


We’re proud to share that the Iguazio MLOps Platform has been named a leader and outperformer in the GigaOm Radar for Data Science Platforms: Pure-Play Specialist and Startup Vendors report. The GigaOm Radar reports take a forward-looking view of the market and are geared towards IT leaders tasked with evaluating solutions with an eye to the future. GigaOm analysts emphasize the value of innovation and differentiation over incumbent market position. In this Radar Report for Data Science Platforms, GigaOm gave Iguazio top scores on several evaluation metrics, including Integrated MLOps Capabilities, Data Security, Scalability, Deployment Flexibility, Usability and Integrations with Third-Party Frameworks. Why GigaOm Named Us an Outperforming Challenger for 2022 Iguazio’s leader placement in the report highlights our differentiated capabilities and our commitment to a production-first approach to delivering AI/ML services. We’re thrilled to see the Iguazio MLOps Platform recognized for the capabilities every company should embrace when adopting MLOps and data science: End-to-end automation, ML CI/CD pipelines, scalability, feature store usage and real-time feature engineering, and more. Of all the vendors reviewed in this report, Iguazio ranked highest for innovation, reflecting the Platform’s aggressive approach to technical innovation, over a more conservative stance. GigaOm noted that nimble startups without legacy technologies are particularly well suited to a rapidly evolving field like data science. GigaOm analysts ranked Iguazio as an outperformer, which denotes their assessment that the Platform will progress rapidly over the next 12 to 18 months, based on strategy and pace of innovation. What the Iguazio MLOps Platform...

---


We are delighted to share that Iguazio has been named along with Microsoft, Databricks, Cloudera, Alteryx and others in Now Tech: AI/ML Platforms, Q1 2022, Forrester’s Overview of the Leading AI/ML Platform Providers, by Mike Gualtieri. This report by Forrester Research looks at AI/ML Platform providers, to help technology executives evaluate and select one based on functionality aligned with their needs. Enterprises who are infusing AI into their applications need a platform that will support the various roles involved. This report provides a survey of the market landscape for technology decision makers. The Forrester report describes Iguazio as a code-first AI/ML platform, for enterprises whose primary teams are coders, along with with GUI tools to perform various tasks around building, deploying and managing models. The code-first paradigm that Forrester identifies in the report allows data scientists and engineers to work together with their preferred open-source tools (like Jupyter), while enabling collaboration via the UI. The Iguazio MLOps Platform is an end-to-end solution for enterprise ML teams, that automates and orchestrates the entire ML pipeline. Data science, data engineering and DevOps team can collaborate on one platform to rapidly deploy operational ML pipelines with an online and offline feature store, built-in model monitoring, dynamic scaling capabilities, all packaged in an open and managed platform. For a complete look at the evaluation criteria and Forrester’s recommendations for navigating this rapidly evolving market, access the full report here (for Forrester subscribers).

---


Machine learning is a practice that is evolving and developing every day. Newfound technologies, inventions and methodologies are being introduced to the community on a daily basis. As ML professionals, we can enrich our knowledge and become better at what we do by constantly learning from each other. But with so many resources out there, it might be overwhelming to choose which ones to stay up-to-date on. So where is the best place to start? We tapped our experienced team to compile a list of eight ML resources for anyone interested in ML: data scientists, data engineers and more. These include YouTube channels, influencers, communities, blogs, and more. We hope you find this list valuable for your professional development. If you have more to add, feel free to drop us a line and we’ll gladly add it to the list. 1. The AI Epiphany A popular YouTube channel by Aleksa Gordić. Each video covers a paper, code or other AI-related topic and breaks it down in a clear and easy-to-follow way. Playlists cover topics like Difusion models, Robotics, Computer Vision, AI breakthroughs, and more. The AI Epiphany has 24. 8 thousand subscribers and more than 600,000 views just two years in the making. Another reason to love this source: Explanations and tutorials are presented in Gordić’s soothing voice and the videos are entertaining. 2. Towards Data Science A Medium-based independent publication that covers all data science-related concepts, like model training, serving, feature engineering, monitoring and more. Consistent reading of articles...

---


When businesses begin applying machine learning (ML) workflows to their use cases, it’s typically a manual and iterative process—each step in the ML workflow is executed until a suitably trained ML model is deployed to production. In practice, the performance of ML models in the real world often degrades over time, as the workflows fail to adapt to changes in the dynamics and data that describe the environment and require frequent retraining with fresh data. Data science teams also need to experiment with new implementations of the ML workflow, such as feature engineering, model architecture, and hyperparameter optimization, to improve ML model performance. This requires experiments tracking to monitor changes in model accuracy that are caused by changes in the ML workflow. In this article, you will learn about the challenges of managing ML workflows, as well as how automating various steps in the workflow using a MLOps approach can help data teams achieve faster deployment of ML models. Understanding the ML Workflow Figure 1: ML workflow with performance monitoring (Source: Adapted from Google Cloud) A typical ML workflow involves the following steps: Data ingestion: Data is extracted from various sources and integrated into an input dataset for the ML task. Data analysis: Data scientists and engineers perform exploratory data analysis to understand the data schema and characteristics, so that they can identify the data preparation and feature-engineering operations needed for building the ML model. Data preparation: Data is prepared from the input dataset for building the ML model. This...

---


Still waiting for ML training to be over? Tired of running experiments manually? Not sure how to reproduce results? Wasting too much of your time on devops and data wrangling? Spending lots of time tinkering around with data science is okay if you’re a hobbyist, but data science models are meant to be incorporated into real business applications. Businesses won’t invest in data science if they don’t see a positive ROI. This calls for the adoption of an “engineered” approach — otherwise it is no more than a glorified science project with data. Engineers use microservices and automated CI/CD (continuous integration and deployment) in modern agile development. You write code, push and it gets tested automatically on some cluster at scale. It then goes into some type of beta/Canary testing phase if it passes the test, and onto production from there.  Kubernetes, a cloud-native cluster orchestrator, is the tool now widely used by developers and devops to build an agile application delivery environment. Leading ML engineers and AI/ML-driven companies are already using Kubernetes, which comes with pre-integrated tools and scalable frameworks for data processing, machine learning and model deployment. Some Kubernetes data pipeline frameworks also enable horizontal scaling and efficient use of GPUs, which further cut down overall wait times and costs. Don’t settle for proprietary SaaS/cloud solutions or legacy architectures like Hadoop. Get a Kubernetes-based solution for a scalable and forward-looking data science platform. The following post reviews best practices for the process and Kubernetes frameworks used to scale and operationalize...

---


Operationalizing AI pipelines is notoriously complex. For deep learning applications, the challenge is even greater, due to the complexities of the types of data involved. Without a holistic view of the pipeline, operationalization can take months, and will require many data science and engineering resources. In this blog post, I'll show you how to move deep learning pipelines from the research environment to production, with minimal effort and without a single line of code. To automate the entire process, we will use MLRun, an open source MLOps orchestration framework. This blog post is based on a talk I gave with Iguazio co-founder and CTO Yaron Haviv at the Open Data Science Conference, called “Automating MLOps for Deep Learning”, which you can watch here. The Challenges of Operationalizing Deep Learning Models  When developing deep learning models, the prevalent mindset most data scientists have today is to begin by developing in the research environment. Data scientists will take the data (images, text, etc. ) from an object, data lake or warehouse, prepare it, train the model and evaluate it. However, this interactive and iterative process quickly breaks down when moved into production. There, it becomes a convoluted mess. This is because the production environment requires the data scientist to make his model a reliable asset in the product, and like any other aspect of the product, the model will require performance monitoring, tests and evaluation, logging of every training/retraining process and versioning in order to pass on the best version of the...

---


The field of MLOps has grown up around the reality that while the theoretical ability of machine learning to make accurate predictions and solve complex problems is incredibly sophisticated, actually operationalizing machine learning is still a major blocker for most companies. Most of the complexities arise from the data: work is typically done in silos, the path to production is resource-intensive, there’s a general lack of simple access to production-ready features at scale that are consistent with the features in the data science research phases, and disjointed or nonexistent model and feature monitoring processes once the AI service is live. ML teams need a way to continuously deploy AI applications in a way that creates real, ongoing business value for the organization.   Features are the fuel driving AI for the organization, and feature stores are the architectural answer that can simplify processes, increase model accuracy and accelerate the path to production. A feature store provides a single pane of glass for sharing all available features across the organization. When a data scientist starts a new project, he or she can go to this catalog and easily find the features they are looking for. But a feature store is not only a data layer; it is also a data transformation service enabling users to manipulate raw data and store it as features ready to be used by any machine learning model. These features can then accelerate machine learning use cases through the reduction of duplicate work. Some of the largest...

---


Hybrid Cloud + On-Premises Model Serving + Model Monitoring Recap Last time in this blog series, we provided an overview of how to leverage the Iguazio Feature Store with Azure ML in part 1. We built out a training workflow that leveraged Iguazio and Azure, trained several models via Azure's AutoML using the data from Iguazio's feature store in part 2. Finally, we downloaded the best models back to Iguazio and logged them using the experiment tracking hooks in part 3. In this final blog, we will: Discuss the benefits of a hybrid cloud architectureDefine model load and predict behaviorCreate a model ensemble using our top three trained modelsEnable real-time enrichment via the feature store during inferencingDeploy our model ensemble in a Jupyter notebook and on a Kubernetes clusterEnable model monitoring and drift detectionView model/feature drift in specialized dashboards Hybrid Cloud Benefits and Motivation Hybrid clouds are all the rage. While cloud computing has given many organizations access to near infinite compute power, the reality is that on-premise infrastructure is not going away. From data privacy concerns, to latency requirements, to simply owning hardware, there are many legitimate reasons for having an on-premise footprint. However, with this increased flexibility comes increased complexity – the right tools are needed for the job. With the multitude of end-to-end platforms, SaaS services, and cloud offerings, it is increasingly difficult to find those correct tools—especially those that work across cloud and on-premise environments. In the last blog of this series, we will combine the...

---


Part 3: Model Training with Azure ML and Iguazio Recap In part one and two, we introduced Iguazio's feature store and discussed the benefits of using one in the ML workflow. Additionally, we ingested and transformed the data that we will be using to train our model. In this blog, we will do the following: Upload data from Iguazio into Azure and register dataset in Azure MLTrain several models in Azure using Auto MLRetrieve trained models from Azure back to IguazioLog trained models with experiment tracking and metrics MLRun Function - Overview Before running any code, we need to take a moment to discuss one of the core tenants of the Iguazio platform: the MLRun function. This abstraction allows for simple containerization and deployment of code. Users are able to execute workloads on Kubernetes by specifying code and configuration options via high level Python syntax. An MLRun function will look something like this: The function will have its own name/tag, code, docker image, resources, and runtime. The syntax for creating an MLRun function looks like the following: From there, the code can be executed locally in Jupyter or on the cluster using several runtime engines including Job, Spark, Dask, Horovod, and Nuclio real-time functions. Create Azure MLRun Function from Python File Now that we have the background on what an MLRun function is, we are going to create one with our Azure code. I have written a Python file called azure_automl. py that performs several tasks such as: Upload the...

---


Part 2: Data Ingestion + Transformation into Iguazio's Feature Store Recap Last time, we discussed why organizations might require the functionality of a feature store like Iguazio's. In this blog, we will actually get into the project and cover the following: Detailed overview of Iguazio feature store functionalityHow to ingest and transform datasets into the feature storeHow to retrieve features in batch, ingest and transform the datasets into feature store, in real time Overview of Iguazio's Feature Store While some feature stores are more focused in their functionality, Iguazio's feature store is designed to facilitate the entire end-to-end workflow: In the previous blog post, we discussed the Iguazio feature store at a high level. The functionalities include (but are not limited to) the following: Ingest and transform data sources in batch or real-timeEasily retrieve features in batch or real-timeDual storage formats to facilitate batch and real-time workloadsComplex real-time feature engineering (e. g. , sliding window aggregations)Integration with model monitoringIntegration with model serving We will be exploring all of this functionality in this blog series. Iguazio Terminology: FeatureSet and FeatureVector The Iguazio feature store introduces two new terms: FeatureSet : group of features from one data source (file, data frame, table, etc. )FeatureVector : group of features from one or more FeatureSets (i. e. a few columns from here, a few columns from there) These act as the building blocks of the Iguazio feature store and will be used heavily throughout this blog series. They fit into the overall picture as...

---


Feature Store Motivation In this series of blog posts, we will showcase an end-to-end hybrid cloud ML workflow using the Iguazio MLOps Platform & Feature Store combined with Azure ML. This blog will be more of an overview of the solution and the types of problems it solves, while the next parts will be a technical deep dive into each step of the process: Part 1: Feature Store MotivationPart 2: Data Ingestion + Transformation into Iguazio's Feature StorePart 3: Model Training via Azure ML leveraging IguazioPart 4: Hybrid Cloud + On-Premise Model Serving + Model Monitoring with Iguazio The Gaps: Challenges When Operationalizing Data Science Regardless of the environment, one of the main challenges when operationalizing data science is fostering collaboration between teams, and eliminating tech silos. This is both a technological and organizational challenge, requiring the right processes in place and the right tools to support these processes. In a typical data science project, there are usually three different teams involved at different points in the pipeline: Data Engineer: Ingest and transform raw data from various sourcesData Scientist: Utilize transformed data to train modelMLOps Engineer: Containerize and deploy model at scale with monitoring, drift detection, and re-training capabilities The Data Science Pipeline - Image by Iguazio While the pipeline itself looks straightforward, there are more than a few places where things can go wrong - mostly at the handoff points between teams. What happens when a data scientist needs additional or different features? What happens when the MLOps Engineer...

---


In an era where the passage of time seems to have changed somehow, it definitely feels strange to already be reflecting on another year gone by. It’s a cliche for a reason–the world definitely feels like it’s moving faster than ever, and in some completely unexpected directions. Sometimes it feels like we’re living in a time lapse when I consider the pace of technological progress I’ve witnessed in just a year. The cool thing about being in the ML industry for so long is that I have a front row seat to a fascinating market characterized by rapid innovation.   So before we toast to a new (and better! ) year ahead, here are my predictions of what awaits the ML industry in 2022: From AutoML to AutoMLOps 2022 will be the start of a focus shift from the practice of model creation to a holistic view towards productizing AI. The next steps will be a set of best practices and repeatable MLOps processes that will help small teams roll out complete AI services on an ongoing basis. We’re not talking about putting a notebook into production, but about building services with ML baked-in, that continuously deliver bottom-line business value.   Until recently, much of the focus was on automating the ML training process, that is, AutoML. But the most significant problem in ML is not about how to find the best algorithm and parameters, but rather about deploying those algorithms as part of an application with business impact. This...

---


After a year packed with one machine learning and data science event after another, it’s clear that there are a few different definitions of the term ‘MLOps’ floating around. One convention uses MLOps to mean the cycle of training an AI model: preparing the data, evaluating, and training the model. This iterative or interactive model often includes AutoML capabilities, and what happens outside the scope of the trained model is not included in this definition.  My preferred definition of MLOps refers to the entire data science process, from ingestion of the data to the actual live application that runs in a business environment and makes an impact at the business level.  This isn’t just a semantic issue--this discrepancy in usage has major consequences for the ways enterprises bring ML to production—or don’t. The Training-First Mindset in MLOps  As a vestige of an era when ML was confined mostly to academia, or to analytics services, most data science solutions and platforms today still start with a research workflow and fail to deliver when it comes time to turn the generated models into real-world AI applications.  So much so, that even the term ‘CI/CD pipeline’ is sometimes used to refer to the training loop, and not extended to include the entire operational pipeline. This mindset forces the ML team to re-engineer the entire flow to fit the production environment.  At this late stage, building the actual AI application, deploying it, and maintaining it in production become acutely painful, and sometimes nonviable. Modern applications in which AI models provide real-time recommendations, prevent fraud, predict failures and guide self-driving cars require significant engineering efforts and a new approach to...

---


Machine learning (ML) model serving refers to the series of steps that allow you to create a service out of a trained model that a system can then ping to receive a relevant prediction output for an end user.   These steps typically involve required pre-processing of the input, a prediction request to the model, and relevant post-processing of the model output to apply business logic.   Out of the vast spectrum of tools and techniques that enable model serving, we will focus on TensorFlow Serving in this blog post. Figure 1: The high-level steps of ML model serving (Source: Kubernetes Blog) Still, serving machine learning models is not a solo act. For a full-scale production system, other external processes perform a vital support role to model serving: services that fetch and ensure consistency of features between serving and training time, services that monitor and log requests, services that automatically deploy the latest model(s), to name a few. If you are solely interested in technical guidance on TF Serving and TensorFlow production deployment, you can jump to the What TensorFlow Serving Can Do for You section for a hands-on guide tailored to different levels of experience. However, if you are also curious about how model serving fits into the intertwined scene of MLOps and how to move toward MLOps maturity, you’ll be interested in what’s coming next. The Need for MLOps and the Central Role of Model Serving The history of machine learning has been marked by a few major...

---


ODSC West Reconnect starts in a few days (Nov. 16-18), which means now is the best time to put the finishing touches to your session attendance schedule. Whether you’re going in-person or joining virtually, ODSC West brings together many AI and data science experts who will present trainings, immersive workshops, keynotes, and more. So if you’re interested in machine learning, data engineering, MLOps, NLP, big data and anything data science - ODSC West Reconnect is the place to be. To help you choose which sessions to attend, we put together a list of our six top recommended sessions. And of course, we will also be there. More details below. Here are our top 6 recommended sessions for ODSC West Reconnect 2021: 1. Unifying Development and Production Environments for Machine Learning Projects Chip Huyen, Adjunct Lecturer, Stanford University Nov. 16, 11:30am - 1:00 pm PT A two-part workshop covering the challenges of productionizing ML models and how to overcome them. The first part of the workshop reviews the topic in theory, while the second part is a hands-on tutorial showing how to use Metaflow to push code to production on AWS batch. (Just make sure you’re comfortable with Python for the tutorial section). This session includes a hands-on tutorial that answers one of the most important MLOps challenges: pushing to production. 2. How to Effectively Scale ML & AI in Any Organization Ella Hilal, PhD, Director of Data Science, Shopify Nov. 16, 12:20 - 12:50 pm, PT This session will cover...

---


You got through all the hurdles getting the data you need; you worked hard training that model, and you are confident it will work. You just need to run it with a more extensive data set, more memory and maybe GPUs.  And then... well.   Running your code at scale and in an environment other than yours can be a nightmare.  You have probably experienced this or read about it in the ML community. How frustrating is that? All your hard work and nothing to show for it. What are the Challenges?   Your coding style will impact how easy it is for others to take your code and run it elsewhere.  For example, hardcoded parameters and file locations can make it difficult to move your code. Those are under your control. You can adopt best practices to make your code more portable. However, you still need to be aware of how your work will integrate with the release process. Is it flexible enough to be part of a CI/CD pipeline?   In today’s enterprise application world, you are dealing with several infrastructure as software technologies:  Docker  Kubernetes  Cloud  Virtual Machines  You can certainly learn all these technologies, but wouldn’t you rather focus on your model training instead? How awesome would it be if you could fully integrate with the infrastructure while concentrating 99% of your time on actual data science with the satisfaction of knowing that your AI/ML project can make it to production fast, at the speed of your business?  This blog will explore how to use MLRun to quickly deploy applications, and run on Kubernetes without changing code or learning a new technology.   Getting Set Up  MLRun can be configured to run on Docker or Kubernetes.  Docker Desktop gives you the most flexibility and faster path to run MLRun.  We are doing this exercise working with MLRun using Kubernetes. You can follow the installation instructions here.   At the end of the installation, you should...

---


By Jean-Baptiste Thomas, Pure Storage & Yaron Haviv, Co-Founder & CTO of Iguazio  You trained and built models using interactive tools over data samples, and are now working on building an application around them to bring tangible value to the business. However, a year later, you find that you have spent an endless amount time and resources, but your application is still not fully operational, or isn’t performing as well as it did in the lab. Don’t worry, you are not alone.  According to industry analysts, that’s the case in more than 85% of AI (artificial intelligence) deployments. Moving from research (using small data sets) to robust online production environments is far from trivial. Production environments need to process large amounts of real-world data and meet application performance and SLA goals. The move from lab to production involves multiple practitioners across teams (data scientists, data engineers, software, DevOps, SecOps etc. ) who need to work in synchronization to build production-grade AI applications.   In order to effectively bring data science to production, enterprises need to change their mindset. Most organizations today initially focus on research and model development because that is often seen as the natural first step to the data science process. However, taking a production-first approach is a sign of operational maturity and strategic thinking. Enterprises that take a production-first approach can plan for the future and make sure to harness the tooling to support the data science process as they mature and later scale. To industrialize AI, an ML (machine learning) solution that delivers scalable production grade AI/ML applications is needed. One that streamlines the workflow and fosters collaboration across teams, with a focus on automation to improve productivity. One option is to build an AI solution from discrete open-source tools, infrastructure components, and/or cloud services, but that requires significant amount of IT and engineering resources. Another option is to choose a pre-integrated and managed solution which lets you focus on solving your...

---


Real-time feature engineering is valuable for a variety of use cases, from service personalization to trade optimization to operational efficiency. It can also be helpful for risk mitigation through fraud prediction, by enabling data scientists and ML engineers to harness real-time data, perform complex calculations in real time and make fast decisions based on fresh data, for example to predict credit card fraud before it occurs. A feature store can help make this process repeatable and reproducible, saving time and effort on building additional AI services. This blog post will explain how real-time feature engineering can be seamlessly performed with an online feature store, and how with seamless integration to an MLOps platform, model monitoring, drift detection and automatic re-triggering of training can be performed, keeping your models at peak accuracy even in changing environments. This article is based on the talk Adi Hirschtein, VP Product at Iguazio, gave at the ML in Finance summit. You can view the entire presentation here.   ML Pipelines in Organizations Companies invest a lot of time and resources putting together a machine learning (ML) team. However, 85% of their projects never make it to production. This is due to business as well as technical challenges. Business Challenges in Building ML Pipelines From the business side, silos between data scientists, ML engineers and DevOps make it challenging for everyone to collaborate on the machine learning pipeline in an efficient way. On top of that, organizations also need to incorporate aspects like security, network, management...

---


With the explosion of the machine learning tooling space, the barrier to entry has never been lower for companies looking to invest in AI initiatives. But enterprise AI in production is still immature. How are companies getting to production and scaling up with machine learning in 2021? Implementing data science at scale used to be an endeavor reserved for the tech giants with their armies of developers and deep pockets. Today, building a machine learning application is feasible for even the leanest startup.   Yet even with the massive growth of the ML tooling space, enterprise AI in production is still immature, lacking a convergence on a common set of best practices and tools. As a response to market conditions caused by the COVID-19 pandemic and its acceleration of all things digital, many companies are forecasting even more investment in AI initiatives in the coming year. To get beyond the experimentation phase, organizations need an automated and streamlined approach to ML operationalization (MLOps). This approach is not just about machine learning workflow automation to accelerate the deployment of ML models to production. It’s also important at an enterprise level, to manage risk as ML gets scaled across the organization to more use cases in dynamic environments, and ensures that the applications continually fulfill business goals.   The role of MLOps to the broader organization shouldn’t be underestimated. Whereas building software is by now a mature and straightforward practice with decades of best practices and a large pool of veteran practitioners,...

---


CPG executives invest billions of dollars in trade and consumer promotion investments every year, spending as much as 15-20% of their total annual revenues on these initiatives. However, studies show that less than 72% of these promotions don’t break even and 59% of them fail. Despite these troubling statistics, most CPG organizations continue to design and execute essentially the same promotions year after year with negligible hope of obtaining sustained ROI.   Trade promotions are becoming increasingly complex and harder to manage as CPG organizations struggle to respond to evolving digital channels of engagement, increasing demands from retailers, and ever-changing consumer behaviors.   Companies are discovering the obsolescence of using manual processes such as basic rules-based technologies and spreadsheets to optimize CPG campaigns. This is due to the proliferation of trade promotions globally, the huge number of SKUs, and constantly evolving channels and consumer preferences. And using these outdated methods to facilitate CPG promotions at scale is virtually impossible.   The billion-dollar question is: How can CPG players successfully run a profitable trade promotion in the face of today’s complex and diverse CPG industry?   Thankfully, emerging technology such as AI and machine learning can help CPG companies execute promotions with greater precision to obtain impactful results and higher ROI. CPG organizations are recognizing the value of AI, ML, and other advanced technologies in enhancing product visibility, increasing brand awareness, and overcoming operational challenges. To watch the full joint webinar with Tredence about MLOps solutions for the CPG industry, check...

---


We are proud to announce that Iguazio has been named a sample vendor in seven 2021 Gartner Hype Cycles, including the Hype Cycle for Data Science and Machine Learning, the Hype Cycle for Artificial intelligence, Analytics and Business Intelligence, Infrastructure Strategies, Customer Experience Analytics, Financial Analytics and Hybrid Infrastructure Services, alongside industry leaders such as Google, IBM and Microsoft (who are also close partners of ours). Iguazio is listed under the following categories: MLOps, Feature Store, AI Orchestration and Automation Platform and Continuous Intelligence. The Hype Cycle for Data Science and Machine Learning, 2021 analyzes how accelerated digitization is driving the urgency to productize experimental data science and machine learning initiatives, and assesses the evolution of existing and emerging trends to orchestrate and productize DSML. The Hype Cycle for Artificial Intelligence, 2021 takes a look at the acceleration of AI as more enterprises embrace digital transformation of their core operations, and how leaders can successfully navigate AI-specific innovations that are in various phases of maturation, adoption and hype. The Hype Cycle for Analytics and Business Intelligence, 2021 evaluates the maturity of innovations across the ABI space, including consumer-focused augmented analytics, composability of D&A ecosystems, and the governance and education required to execute a variety of analytics at scale. The Hype Cycle for Infrastructure Strategies, 2021 covers innovations and enhancements in infrastructure consumption models, automation/intelligence and architecture. The report includes a look at net zero data centers, the disruptions and opportunities of containers and cloud delivery, and the maturity of software-defined...

---


We just wrapped up the first-ever 'MLOps for Good' hackathon, and we are so thrilled by the incredible response we’ve gotten from the ML community.   300 participants from all parts of the globe—From the USA, Canada, Germany, Singapore, New Zealand, Australia, India, Portugal, Philippines, Malaysia, Morocco and Pakistan—joined our cause of bringing data science to production for social good. After 6 weeks of intense hacking, 30 innovative projects were submitted. It is gratifying to pause and celebrate the ingenuity these hackathoners have shown. Last night we held the virtual Finals and Awards Ceremony. At the ceremony, our global audience heard from our panel of expert judges, including: Cecile Blilious - Head of Impact and Sustainability at Pitango Venture Capital Greg Hayes – Data Science Director, Ecolab Orit Nissan-Messing - Co-Founder and Chief Architect, Iguazio Nick Brown – Senior Data Scientist, IHS Markit Anna Anisin - Founder, DataScience. Salon Yaron Haviv – Co-Founder and CTO, Iguazio Adi Hirschtein - VP Product, Iguazio They also heard from our keynote speakers, Tomer Simon, Chief Scientist at Microsoft Israel R&D Center, and Boris Bialik, Global Head of Enterprise Modernisation at MongoDB. Simon leads the Microsoft AI for Good program, and shared how Microsoft is donating resources and technology to solve global challenges relating to the environment, accessibility, healthcare, cultural heritage, and humanitarian issues. The goal is to “amplify human ingenuity, and bring it to scale ” he said. He encouraged the hackathon participants to check out the grants available with the AI for...

---


---


## News


---


## Events


MLOps NYC 2019 gathered industry leaders from companies like Netflix, Google, Twitter and Uber to share their biggest MLOps challenges and solutions.

---


The Cloud Native Computing Foundation’s flagship conference gathers adopters and technologists from leading open source and cloud-native communities in Amsterdam

---


---


## PRs


Acquisition will enable Artificial Intelligence’s (AI) full power and potential to be realized across commercial, social, and environmental initiatives. McKinsey & Company today (23rd January) announced that it has acquired Iguazio, a Tel-Aviv-based leader in Artificial Intelligence and Machine Learning. McKinsey will be able to accelerate and scale AI deployments dramatically with the addition of Iguazio’s technology and a team of 70+ data and AI experts. To thrive in today’s competitive market, harnessing the power of Artificial Intelligence (AI) is essential. According to McKinsey research, more than $490B was invested in AI by organizations around the globe from 2012-2021. But for most, the actual value of those investments has yet to be realized, with only one in ten projects making it outside the lab. QuantumBlack, AI by McKinsey has been working with clients to address these challenges head-on and embed AI into real-time decision-making. The final element we have been working on is a technology solution that will accelerate AI deployment, embedding it in real-time and in any environment. “After analysing more than 1000 AI companies worldwide, Iguazio was identified as the best fit to help us significantly accelerate our AI offering – from the initial concept to production, in a simplified, scalable and automated manner,” said Ben Ellencweig, McKinsey senior partner and QuantumBlack global leader of alliances and acquisitions. “By joining forces with Iguazio, we can now deepen the unparalleled, disruptive, end-to-end AI capabilities we offer to our clients. ” Working with Iguazio, QuantumBlack will now be able to...

---


Sense will use the Iguazio MLOps platform for a large range of AI products, beginning with the Sense AI Chatbot, an intelligent, automated recruiting assistant that speeds the hiring process and provides uncompromised personalization, already available to 700+ customers.    San Francisco, 27th July, 2022 – Iguazio, the MLOps platform provider, and Sense, a market leader in artificial intelligence (AI) driven talent engagement for recruiting, today announced that the Iguazio MLOps platform has been selected to power a wide range of AI products aimed at increasing the efficiency and scalability of Sense’s AI operations. Sense is leveraging automation and AI to speed up the recruitment process, while delivering a hyper-personalized candidate experience. Sense is a leader in talent engagement, with over 700 customers and an annual growth rate of over 100%. Sense provides the Sense AI Chatbot - an automated recruiting assistant that can engage with candidates 24/7, responding to their queries in real-time even when human recruiters are offline. It engages with candidates across SMS, mobile, and web, matches them to jobs, schedules interviews, and handles intelligent communications, including FAQs. The chatbot pairs conversational AI with automated communication and engagement workflows so organizations can engage with candidates at scale. Sense has a large team of data scientists and machine learning (ML) engineers with deep expertise in conversational AI – both voice and text. The team’s challenge was building the complex natural language processing ( NLP) serving pipeline, with custom model ensembles, to track question-to-question context and enable sentiment tracking....

---


The Iguazio MLOps Platform and built-in Feature Store now offers connectivity to the Snowflake Data Cloud, providing a full solution for enterprises looking to generate real business value with data science in an efficient, scalable and repeatable way Las Vegas, 14th of June, 2022 – Iguazio, the MLOps platform provider, today announced a new partnership with Snowflake, the Data Cloud company, which includes connectivity of Iguazio’s solution for automating ML pipelines and built-in feature store with Snowflake’s Data Cloud. The solution is already being implemented by numerous customers, including Fortune 500 companies, to abstract away the complexities of MLOps and create a joint, production-ready environment for data scientists to continuously roll out new AI services across the organization. The Iguazio MLOps platform can accelerate the data science process up to 12x and make more efficient use of AI resources, like GPUs, through better orchestration and automation.   “We are excited to partner with Iguazio to bring end-to-end MLOps capabilities, including a built-in feature store, to our joint customers,” said Tarik Dwiek, Sr. Director Technology Alliances at Snowflake. “We see the need for MLOps automation continuing to grow as data science matures in the enterprise and we look forward to offering deeper integrations with Iguazio in the coming months. ” The Iguazio platform now comes with a built-in Snowflake connector that powers the built-in online and offline feature store in Iguazio with data from Snowflake, allowing enterprises to seamlessly access the Data Cloud to build, store and share features that are...

---


LATAM Airlines Group is planning to deploy over 40 AI products on the Iguazio MLOps platform in the upcoming weeks, paving the way for other airlines that are choosing AI innovation as their strategy for 2022  NEW YORK, NY, 29th of March, 2022 – Iguazio, the MLOps (machine learning operations) platform provider, today announced that LATAM - the leading airline group in Latin America - has selected its MLOps platform for a large scale, cross-company AI innovation project. The project will span the entire organization and will include use cases such as optimizing and safeguarding the company’s popular frequent flyer program from fraud, improving pilot training through better understanding of the factors that create un-stabilized approaches to landing, and intelligent route planning to reduce CO2 emissions. “The airline industry needs to rethink its strategy post-COVID”, commented Juliana Rios, IT & Digital Vice President at LATAM. “We chose the Iguazio MLOps platform to operationalize data science, fueling more efficient and environmentally-friendly operations, enhanced safety, and better customer service”. LATAM Airlines Group works extensively with GCP, utilizing tools like Google Big Query, Google Cloud Storage, and Google Workload Identity. Iguazio is fully compatible with GCP and has a strong partnership with Google. LATAM Airlines is planning to deploy its AI products on GCP using Iguazio. “LATAM is leading the way for other Airlines in their vision, and how they are using AI to create business value and plan for the post-pandemic future”, said Asaf Somekh, Co-Founder and CEO at Iguazio. “They have...

---


Through integrations with Pure Storage and the Iguazio MLOps platform, enterprises can deploy and scale AI across multi-cloud and hybrid use cases  New York, NY, 26th October 2021 - Iguazio, the MLOps (machine learning operations) company, today announced that it has become a Pure Storage technology partner. The new partnership will empower enterprises to unlock the value of their data and bring data science projects to life in an efficient, automated, and repeatable way. Together, Iguazio and Pure Storage empower enterprises to continuously roll out new AI services by adopting a production-first mindset using technologies that later allow them to scale with AI. By providing enterprise-grade capabilities around data management, scale and performance, and a layer of abstraction and automation, enterprises can start to focus on their business applications and not the underlying infrastructure.   “Many of our customers are seeing the enormous potential for AI-driven innovation, so we strive to provide our customers with the tools and infrastructure they need for the next step of AI” said Grace Chung, Director, Strategic Alliances, Analytics at Pure Storage.  “We’re excited to help Pure and Iguazio’s join customers unlock the value of their data with an agile MLOps solution where compute and storage scales along with the needs of the business”.   “Most organizations starting out with AI initially focus on research and model development, because that is often seen as the natural first step to the data science process”.  Commented Asaf, Co-Founder and CEO of Iguazio.  “However, in order to effectively industrialize AI, enterprises need to think about the components they will need once they are running their AI applications in production at scale - as early on as possible. Together, Pure and Iguazio provide a complete MLOps solution with enterprise-grade data management capabilities, providing the necessary building blocks to efficiently and cost-effectively support diverse and growing AI use cases which fuel business growth”. With the new partnership, Iguazio offers an MLOps platform...

---


Iguazio is the first MLOps platform to enable FSx for ONTAP as a part of its end-to-end capabilities for bringing data science to production at scale, in real-time and in hybrid environments NEW YORK, NY, 20th October, 2021 – Iguazio, the MLOps (machine learning operations) company, today announced its support for the new FSx for ONTAP. FSx for ONTAP provides fully managed shared file and block storage on AWS Cloud with the popular data access and management capabilities of ONTAP.   Iguazio provides a leading MLOps platform used by enterprises worldwide to accelerate the deployment of AI in production, reduce complexities and minimize cost of AI infrastructure through end-to-end ML pipeline automation. Iguazio also supports complex use cases requiring scale and real-time, or deployment in hybrid environments. “So many of our customers are adopting cloud native strategies and AI to create new competitive advantages, increase agility and reduce costs”, commented Ronen Schwartz, SVP and GM Cloud Volumes at NetApp. “Enterprises building AI Applications at scale with enterprise grade storage now have the option to efficiently roll out new AI services and support real-time ML applications at peak performance with Iguazio and FSx for ONTAP, all from within their AWS account. Iguazio is a strategic NetApp partner. The new support of FSx for ONTAP provides customers with enterprise-level data management, advanced storage services including as tiering and snapshotting, and high performance and scalability, supporting the most extreme workloads. It provides a simple end-to-end solution for deploying and managing AI applications at...

---


AWS customers globally can now access Iguazio and unlock a faster path to the deployment of AI at scale NEW YORK, NY, 6th of October, 2021 – Iguazio, the MLOps (machine learning operations) company, today announced its availability in the AWS Marketplace, a digital catalog with thousands of software listings from independent software vendors that makes it easy to find, test, buy, and deploy software that runs on Amazon Web Services (AWS). This new availability provides AWS customers with access to Iguazio’s MLOps solution, which automates machine learning (ML) pipelines end-to-end and accelerates deployment of artificial intelligence (AI) to production by 12X. By way of example, the Hydroinformatics Institute in Singapore (H2i) uses Iguazio on AWS to build and run a real-time ML pipeline that predicts rainfall by analyzing videos of cloud formations and running CCTV-based rainfall measurements. “With Iguazio, we are now able to analyze terabytes of video footage in real time, running complex deep learning models in production to predict rainfall,” said Gerard Pijcke, Chief Consultancy Officer, H2i. “Repurposing CCTV-acquired video footage into rainfall intensity can be used to generate spatially distributed rainfall forecasts leading to better management of urban flooding risks in densely populated Singapore. ” AWS customers can now purchase Iguazio through their AWS Marketplace account in just a few clicks, without having to deal with new contracts or legal hassles. “Iguazio is a great addition to AWS Marketplace, with their unique solution to accelerate AI deployment, even at scale, in real time,” said Mona Chadha,...

---


Tel Aviv and London, 15th June 2021 - Iguazio, the data science & MLOps platform build for production, today announced a strategic partnership with Boston Limited, an NVIDIA Elite Partner and leading provider of high-performance, mission critical server and storage solutions. The partnership enables both companies to extend their offerings to enterprises across industries looking to bring data science into real life applications, and accelerate their path to production. Data science is becoming a critical element of business strategy in enterprises across industries. Companies need better ways to implement their AI solutions in real-world environments, to help them cut costs, work more efficiently, and accelerate the rollout of new AI services and products for customers. Yet as enterprises navigate the journey from data science to live AI applications, they often find the move to production challenging and complex. They need new tools and technologies to help manage this, especially as they scale. This radical shift to transforming business models with AI typically requires highly customizable infrastructure, as well as a streamlined data science workflow to navigate the transformation effectively and efficiently. With the new partnership, Boston Limited will offer high-performance data center hardware and technical services, while Iguazio provides its data science platform which saves time and cost on getting AI to production. “Our primary focus is to provide our customers with the ability to customize their solutions based on their toughest business requirements. The partnership with Iguazio allows us to facilitate greater enterprise AI capabilities within our existing storage...

---


Data Scientists, Data Engineers and MLOps Practitioners called upon to bring data science to production for immediate social impact. NEW YORK, NY, May 24, 2021 --Iguazio, the Data Science Platform built for production, today announced its first-ever global virtual hackathon, which is starting today and will take place until June 29th, 2021. With a mission to foster projects that can immediately impact real-world issues, Iguazio, partners Microsoft, MongoDB, and sponsor Aztek seek individuals who want to bring data science to production to do good together. Teams can register to join the hackathon at mlopsforgood. devpost. com. 2020 has been a tough year. AI models that can tackle topics like improving healthcare, detecting fake news and making the web safer for children have the potential to create positive change in many areas, but unfortunately, bringing data science to production so that it can generate real-world impact is still a largely unsolved challenge. This hackathon calls out to data scientists, data engineers and MLOps practitioners who want to create real world impact now, by building not models but fully functional AI applications to generate immediate positive change. Global challenges are defined by the UN Sustainable Development Goals and will serve as guidestars for the teams to focus on real-life social and environmental goals. Submissions will be judged on solution innovation, commerciabilty & applicability, and ML operationalization / repeatability or the completeness of the functional ML pipeline, with additional points awarded for building real-time pipelines, creating the ability to retrain models, or incorporating...

---


Iguazio Launches the First Integrated Feature Store within its Data Science Platform to Accelerate Deployment of AI in Any Cloud Environment The first production-ready integrated solution for enterprises to catalogue, store and share features centrally, and use them to develop, deploy and manage AI applications across hybrid multi-cloud environments Iguazio’s feature store tackles one of the greatest challenges in machine learning operations (MLOps) today - feature engineering The feature store is a key component in Iguazio’s data science platform, which is used by customers such as Payoneer, Quadient and Tulipan to deploy AI faster, and has just been selected by the Sheba Medical Center to deliver real-time AI for COVID-19 patient treatment optimization Joint solutions with strategic partners Tredence, NetApp, MongoDB and others are already enabled by Iguazio’s feature store, offering reproducible real-time ML pipelines Iguazio Founders left to right: Yaron Haviv, Yaron Segev, Orit Nissan-Messing, Asaf Somekh NEW YORK, NY, December 16, 2020 --Iguazio, the Data Science Platform built for production and real-time machine learning (ML) applications, today announced that it has launched the first production-ready integrated feature store. The feature store, which sits at the heart of its data science platform, enables enterprises to catalogue, store and share features for development and deployment of AI in hybrid multi-cloud environments and is built to handle real-time use cases. According to Gartner, one of the top barriers to AI implementation is the “complexity of AI solution(s) integrating with existing infrastructure”. At the core of machine learning is the data, and...

---


Iguazio was selected to facilitate Sheba’s transformation with AI through clinical and logistical use cases such as predicting and mitigating COVID-19 patient deterioration and optimizing patient journey with smart mobility Joint projects include collaboration on hybrid and multi-cloud AI deployments using Microsoft Azure and Google GCP On Dec. 30, Sheba will be holding a Big Data and AI conference, where the projects will be presented. Other hospitals and medical centers worldwide are invited to get in touch with ARC (Accelerate Redesign Collaborate) at Sheba for more information and to discuss how to implement real-time AI in their health facilities Sheba ARC and Iguazio Ink Agreement from left to right: Asaf Somekh, Eyal Zimlichman (Herzliya, Israel, December 15, 2020) -- Iguazio, developers of the Data Science Platform built for production and real-time machine learning (ML) applications, announced that it is working with the Sheba Medical Center’s ARC innovation complex to deliver real-time AI across a variety of clinical and logistical use cases in order to improve COVID-19 patient treatment. Sheba is the largest medical facility in Israel and the Middle East and has been ranked amongst the Top 10 Hospitals in the World by Newsweek magazine. Iguazio was selected to facilitate Sheba’s transformation with real-time AI and MLOps (machine learning operations) in a variety of projects. One of these projects is optimization of patient care through clinical, real-time predictive insights. Using the Iguazio Data Science Platform, Sheba is incorporating real-time vital signs from patients by utilizing the patient’s medical history to...

---


NEW YORK, NY, November 30, 2020 --Iguazio, the Data Science Platform built for production and real-time machine learning (ML) applications, today announced that it has achieved the AWS Outposts Ready designation, part of the Amazon Web Services (AWS) Service Ready Program. This is a notable development for AWS and Iguazio customers who can utilize Amazon SageMaker to develop artificial intelligence (AI) models and data pipelines, and easily deploy and manage these in production using the Iguazio Data Science Platform on AWS and now also on AWS Outposts, benefiting from the same high performance at scale in hybrid AWS environments.   Two main challenges are hindering adoption of AI for Enterprises and Government agencies. The first is an increase in the need for hybrid solutions to manage data and data science applications, to address data locality in accordance with a rise in regulation and data privacy considerations. The second is an increase in first-hand experiences with the challenges and complexities involved in operationalizing machine learning, especially when considering hybrid deployment options, and when scaling data science across the organization. AWS Outposts is a fully managed service that extends AWS infrastructure, AWS services, API and tools to virtually any datacenter, co-location space, or on premises facility for a truly consistent hybrid experience. Iguazio’s platform offers a fully automated MLOps solution for data engineers, data scientists, and DevOps teams, which includes a high-performance serverless framework and a fast online and offline feature store. The new seamless integration of Iguazio’s technology with AWS Outposts...

---


PadSquad uses Iguazio to deliver high-performing interactive campaigns served in real-time for Fortune 500 brands including Intel, Verizon and Novartis. The company analyzes behavioral data from multiple sources in order to serve the right ad creative to the right customer at the right moment. Iguazio selected for ease of AI deployment and ability to bring data science solutions to market quickly and cost-effectively.   Padsquad uses Iguazio’s platform end-to-end on AWS. NEW YORK, NY, September 21st, 2020 --Iguazio, the data science platform built for production and real-time machine learning applications, today announced it has been deployed by mobile software company PadSquad, to improve the relevance and performance of the digital campaigns they run for their customers worldwide. PadSquad is revolutionizing traditional media with interactive features and innovative technologies that transform the audiences’ experience and engagement with ad creatives. Iguazio was deployed by PadSquad to use AI to improve ad performance and reduce media costs for their customers. They do this by ingesting and acting upon real-time events - from contextual content on the page, engagement with creative elements like video views, swipeable panels and hot spots, to the season and time of day - at a rate of over 3,000 events per second. Utilizing online and offline behavioral data from multiple sources, available to them through third party platforms and their own internal tools, Padsquad can now harness machine learning to optimize ad performance and provide a better and more personalized user experience for their customers’ audiences. “Real time AI...

---


HERTZELIA, Israel, 21 September 2020 - Iguazio, the data science platform built for production and real-time machine learning applications, today announced a strategic partnership with Faktion, a boutique, end-to-end Artificial Intelligence (AI) service provider for smart mobility technologies across Europe.   The partnership enables both companies to provide AI software infrastructure and services to smart mobility companies looking to harness big data and create AI applications that make travel in cities safer, greener, and more efficient. The two companies already have several joint European customers in the Smart Mobility space, tackling challenges such as dynamic congestion charges based on real-time data and detecting and acting upon driver fatigue.   Congestion cost the U. K. nearly £8 billion in 2018. This is the main problem Airvi, a startup which works with cities to regulate and reduce traffic congestion and pollution according to data delivered in real-time, is working to fix. Airvi works with cities to regulate and reduce traffic congestion and pollution using data delivered in real-time. Along with Faktion and using Iguazio's Data Science Platform, it is creating virtual, dynamic zones that adapt to fluid air and congestion patterns to better manage traffic. By harnessing data at scale, from sources such as cell phones, weather sensors and other diverse data sources, through Iguazio’s platform, it can recommend dynamic tariffs - that will deter drivers from entering high traffic and pollution zones.   “Faktion’s deep industry experience and Machine Learning skill set, combined with Iguazio’s platform, has made our product viable...

---


● Top tier consultancy partners with a leading data science platform company to simplify and expedite development and deployment of AI for enterprises across industries: Finance, insurance, healthcare, retail, manufacturing, gaming, AdTech, etc. ● Enterprises can now easily incorporate AI and MLOps automation to create business impact through endless applications such as: Predictive maintenance, real time recommendations, KYC (know your client) and fraud prevention● The partnership will speed up deployment of AI services at lower cost 14 July 2020 – Tel Aviv and New York, NY - Iguazio, the data science platform for real-time machine learning applications, today announced a strategic partnership with SFL Scientific, a leading data science consulting firm. The partnership will enable both companies to extend their offerings to enterprises of all industries looking to apply AI to real life applications, regardless of the size or skill set of their internal teams. In today’s economic environment, enterprises across industries are looking to develop viable AI solutions that help them cut costs, work more efficiently, and develop new services and products for customers. However, as enterprises navigate the journey to develop viable AI solutions and derive the business benefits from analyzing big data, they must transform not only their workforce, standard processes, and operating models, but also modernize critical applications in their infrastructure and architecture. This is often a tremendous task, and one that requires expert support to navigate the transformation effectively and efficiently. With the new partnership, SFL Scientific will offer data strategy and support for algorithm...

---


NetApp deploys the Iguazio Data Science Platform to boost the infrastructure behind its Active IQ solution, responding in real-time to 10 trillion data points per month collected from storage controllers globally, providing actionable intelligence for predictive maintenance and optimal data management. New York, June 10th, 2020 – Iguazio, the data science platform for real-time machine learning applications, today announced a new strategic customer, NetApp, who is using its platform to analyze 10 trillion data points per month, to automate the support and optimization of storage. NetApp’s Active IQ uses predictive analytics to automate the proactive care and optimization of storage controllers owned by customers around the world. NetApp wanted to build a digital advisor that uses AI at scale and in real-time to continually gain insights on these devices and conduct predictive maintenance on storage, while constantly learning and getting smarter over time. Iguazio has implemented similar predictive maintenance solutions for other customers in the past. Previously built on Hadoop, NetApp was also looking to modernize the service infrastructure to reduce the complexities of deploying new AI services and the costs of running large-scale analytics. In addition, the shift was needed to enable real-time predictive AI and to abstract deployment, allowing the technology to run on multi-cloud or on-premises seamlessly. NetApp turned to Iguazio to replace their traditional data warehouse and Hadoop-based data lake with a Kubernetes-powered, cloud-native, serverless data science platform which can analyze massive amounts of data in real-time. The platform is deployed both in the cloud and...

---


The Iguazio Data Science Platform enables greater GPU efficiency and helps to democratize AI infrastructure deployment for every enterprise. New York, May 14, 2020 - Iguazio, the data science platform for real-time machine learning applications, announced today that the company’s solution for utilizing GPU-as-a-Service has been certified as part of the NVIDIA DGX-Ready Software program. Managing and controlling resources as part of ML pipelines, which span data processing, model training, and inference, can pose challenges such as inefficient resource sharing, workload and data scaling, and high development and operations overhead. The Iguazio Data Science Platform enables data scientists and data engineers to work in one self-service data science platform, which dynamically allocates GPUs to workloads for greater resource efficiency. The solution supports horizontal scaling, helping teams address bigger and more complex tasks more quickly, and automates end-to-end ML pipelines from data collection to model creation. The Iguazio platform is based on the leading open source frameworks in a hardened and ready to use package. It also includes the ability to monitor the entire stack from low-level GPU usage, through ML jobs and data, to end-to-end pipelines — simplifying the operations of GPU-based systems. In addition, Iguazio’s open source projects, Nuclio and MLRun, are the only serverless and orchestration frameworks that address data and ML workloads, by optimally managing GPU resources, and supporting stateful and distributed workloads, combining simplicity, faster performance, and costs savings. The GPU-as-a-Service solution from Iguazio is now part of the NVIDIA DGX-Ready Software program, which delivers proven...

---


New York, May 4th, 2020 – Iguazio, the data science platform for real-time machine learning applications, today announced a strategic partnership with NetApp that provides enterprises with a simple, end-to-end solution for developing, deploying and managing AI applications at scale and in real-time on top of the ONTAP AI framework. Despite the great promise of AI for business applications, many data science projects fail to create business value. In fact, according to Gartner, 85 percent of data science projects fall short of expectations. One of the reasons is that model creation is just the first step, while moving working models into a production environment introduces a whole set of complexities, such as handling data at scale, working in hybrid environments, and harnessing real-time data for predictive applications. Businesses can overcome these challenges by working in a production-ready environment, with a simplified infrastructure that enables them to focus on creating business value. Iguazio’s data science platform provides end-to-end machine learning pipeline automation, coupled with performance & scale, enabling real-time machine learning (ML) applications. It introduces a fresh approach to simplifying MLOps and enabling enterprises to deploy their AI projects quickly and seamlessly. Iguazio’s integration with NetApp ONTAP AI leverages enterprise grade data management, data versioning and NetApp Cloud Volumes for a seamless hybrid cloud experience. It is also fully compatible with KubeFlow 1. 0, offering a managed KubeFlow solution for enterprises. The platform tightly integrates with NVIDIA DGX, allowing customers to utilize GPU-as-a-Service and NGC containers, making more efficient use of...

---


Payoneer uses Iguazio to move from detection to prevention of fraud prevention with predictive machine learning models served in real-time New York, January 13th, 2020 - Iguazio, the data science platform for real time machine learning applications, today announced that Payoneer, the digital payment platform empowering businesses around the world to grow globally, has selected Iguazio’s data science platform to provide its 4 million customers with a safer payment experience. By deploying Iguazio, Payoneer moved from a reactive fraud detection method to proactive prevention with real-time machine learning and predictive analytics. Payoneer overcomes the challenge of detecting fraud within complex networks with sophisticated algorithms tracking multiple parameters, including account creation times and name changes. However, prior to using Iguazio, fraud was detected retroactively, enabling customers to only block users after damage had already been done. Payoneer is now able to take the same sophisticated machine learning models built offline and serve them in real-time against fresh data. This ensures immediate prevention of fraud and money laundering with predictive machine learning models identifying suspicious patterns continuously. The cooperation was facilitated by Innovigates-BeLocal, one of the leading Data and IT solutions integrator for mid and enterprise companies. “We’ve tackled one of our most elusive challenges with real-time predictive models, making fraud attacks almost impossible on Payoneer” noted Yaron Weiss, VP Corporate Security and Global IT Operations (CISO) at Payoneer. “With Iguazio’s Data Science Platform, we built a scalable and reliable system which adapts to new threats and enables us to prevent fraud...

---


Iguazio’s data science platform automates machine learning pipelines, enabling a wide range of industries to bring their data science to life. This investment brings Iguazio’s total funding to $72M. HERZLIYA, Israel – January 27th, 2020 – Iguazio, the data science platform for real time machine learning applications, today announced that it has raised $24M of funding. The round was led by INCapital Ventures, with participation from existing and new investors, including Samsung SDS, Kensington Capital Partners, Plaza Ventures and Silverton Capital Ventures. The funds will be used by Iguazio to accelerate its growth and expand the reach of its data science platform to new global markets. The demand for AI applications is on the rise. According to Gartner, AI augmentation alone will create $2. 9 trillion of business value in 2021. However, there are still many challenges in deploying AI solutions in an effective and scalable way. An estimated 87% of data science models which have shown great promise in the lab never make it to production. This is due to the challenges of transforming a great AI model, which is functional in lab conditions, to a fully operational AI application that can deliver business impact at scale and in real time. Iguazio solves this problem and brings data science to life for enterprises worldwide. The Iguazio data science platform helps data scientists create real-time AI applications while working within the familiar machine learning stack they know and love. The platform has been deployed by enterprises spanning a variety of...

---


PICSIX uses Iguazio to provide an AI-based platform addressing the ever-changing threats to homeland security and public safety Paris, Milipol 2019, November 19th 2019 – Iguazio, the Data Science Platform for automating machine learning pipelines, today announced that PICSIX, a leader in tactical intelligence solutions is using its platform to deliver an AI-based Investigative Intelligence Platform. The platform makes real-time machine learning accessible to any agency, providing the flexibility required to address a diverse range of ever-changing threats. As opposed to other managed solutions, PICSIX customers design their own mission specific workflow, at a substantially lower cost. Furthermore, with Iguazio under the hood, users work in an efficient and open environment which automates technical heavy lifting and easily integrates with third-party systems and open source tools. Iguazio’s unified data layer analyzes all types of data sources at scale and in real-time, powering the integration of PICSIX Tactical Intelligence databases, government databases and publicly available electronic information (PAEI) used for OSINT. Machine learning models are applied on the data, generating up-to-date profiles and interactive dashboards for real-time insights, alerts and actions. PICSIX deploys the system in multiple edge locations, while Iguazio’s Data Science Platform can also be deployed on-premises or in multi-cloud environments. “PICSIX has been aiding law enforcement and homeland security agencies in their continuous battle against terror, drug trafficking, human trafficking and more, with the ultimate goal of making the world a safer place,” said Menachem Kenan, CEO. “By deploying our Intelligence Investigative Platform on Iguazio, we are able...

---


New serverless capabilities in Iguazio’s Data Science Platform enable on-demand resource consumption, elastic scaling, and simpler ML pipelines New York City, MLOps NYC19, September 24th 2019 – Iguazio, the Data Science Platform for automating machine learning pipelines, today announced Nuclio ML Functions, broadening the serverless capabilities of Iguazio’s Data Science Platform for scalable machine learning training and data preparation. Nuclio is the only open source serverless framework that extends beyond event driven workloads to long lasting, parallel, and data-intensive jobs. This brings the benefits of serverless - allocating resources on demand, auto-scaling, and dev ops automation - to machine learning, data preparation and analytics workloads. Serverless frameworks have so far only addressed challenges posed by the beginning and end of ML pipelines, namely data ingestion and model serving. Iguazio’s open source Nuclio has stood out with its unmatched performance and parallelism, its support of GPUs at scale, multi-cloud deployment and native integration into the data science stack. With Nuclio ML Functions, Iguazio now provides a layer of automation and monitoring for the widely used ML and analytics frameworks on top of Kubernetes, relying on its high-speed shared data layer for seamless scaling. ML function activities and data artifacts are automatically logged, allowing users to trace data and experiment results and re-run older jobs when needed. Nuclio ML Functions also leverages Kubeflow to speed up the running of ML pipelines. With the introduction of these new serverless features, Iguazio enables full automation and CI/CD for ML workloads, cutting infrastructure costs, tedious...

---


Speakers from Google, Walmart, Netflix, Uber, Twitter, Microsoft, Bloomberg and more will convene in New York to improve machine learning automation of development and deployment New York City, August 6th 2019 – Iguazio, the Data Science Platform for automating machine learning pipelines, today announced MLOps NYC19, a community-organized conference taking place September 24th in New York’s Hudson Mercantile. The MLOps Call for Proposals is open and early bird ticket sales end August 24th. More than half of data science projects are not fully deployed, according to Gartner: “Many organizations struggle when it comes to systematically productizing machine learning results, as the production process is either overlooked or left solely to the DevOps team. ”* MLOps NYC19 will reflect the current state of machine learning operations with accomplished industry leaders sharing insights and experiences, such as Bill Groves (Walmart), Julie Pitt (Netflix), Karl Weinmester (Google), Brittany Wills (Twitter) and Josh Patterson (NVIDIA). Attending MLOps NYC19 are data scientists, machine learning engineers and enterprise CTOs and CDOs, to participate in presentations and training sessions about: AI in business applications Kubeflow and MLSpec standardization MLflow Serverless in machine learning Best practices for ML workloads ML workflows for GPUs “Machine learning is expanding from research and bleeding edge companies to any modern app,” said Yaron Haviv, Iguazio CTO. “Much like other software practices, the industry needs to define and adopt successful ML development and CI/CD patterns. MLOps NYC19 will drive collaboration to accelerate mainstream ML adoption, enabling portability and interoperability. ” Microsoft’s Head of...

---


Enabling Leading Enterprises to Bring Machine Learning into Business Applications and Remove AI Project Complexity Microsoft Build Seattle, May 8th 2019 – Iguazio, provider of the data science platform built for production, today announced the availability of its platform on Microsoft Azure and Azure Stack, serving a wide range of cloud and intelligent edge use cases. The platform enables fast development and deployment of data science applications with unified data services, serverless functions and integrated AI tools, all consumed in a managed environment. Iguazio is a Microsoft co-sell partner and offers its data science platform through the Azure Marketplace. “Iguazio is proud to collaborate with Microsoft to enable AI-driven actions both with Microsoft Azure and Azure Stack at the intelligent edge,” said Iguazio CEO, Asaf Somekh. “Iguazio’s platform brings data science to life with its production-native architecture and it can now serve the entire data science lifecycle with Azure and Azure Stack. Combining Iguazio and Azure, harnessing AI and cloud, enables high paced innovation and the faster delivery of advanced services. ” Henry Jerez, Principal Group Product Manager at Microsoft’s Intelligent Edge Solutions Platform Group said, “Partnering with Iguazio we can offer additional options for AI applications in the cloud to also run on the edge. Iguazio provides an additional path to run AI on the edge beyond our current Microsoft Azure Machine Learning inferencing on the edge. This new marketplace option provides an additional alternate path for our customers to bring intelligence close to the data sources for applications...

---


Samsung SDS Integrates Serverless and Big Data to Automate Machine Learning Application Scaling and Productization GTC San Jose, March 19th 2019 – Iguazio, provider of the high-performance platform for serverless and machine learning applications, today unveiled native integration with NVIDIAⓇ GPUs to eliminate data bottlenecks, provide greater scalability and shorten time to production. Iguazio’s platform powers machine learning and data science over Kubernetes, enabling automatic scaling to multiple NVIDIA GPU servers and rapid processing of hundreds of terabytes of data. Iguazio’s data science platform provides: Serverless functions that run data-intensive applications on NVIDIA GPU-enabled servers. Integration of its database with RAPIDS, NVIDIA’s open-source machine learning libraries, for faster and scalable data processing. The use of Iguazio’s serverless functions (Nuclio) improves GPU utilization and sharing, resulting in almost four times faster application performance when compared to the use of GPUs within monolithic architectures. Nuclio is fifty times faster than serverless solutions that do not offer GPU support, such as Amazon’s Lambda. Serverless and Kubernetes target key challenges in data science: they simplify operationalization, eliminate manual devops processes and cut time to market. Samsung SDS will use Iguazio to speed up machine learning applications and leverage the automatic scaling of Iguazio’s serverless framework to increase efficiency and sharing. Samsung SDS announced its investment in Iguazio on March 6th. The company has accelerated its pipeline with Iguazio to streamline the delivery of intelligent applications, analyzing models built directly in its production environment and generating predictions. “The integration of Iguazio with NVIDIA RAPIDS provides...

---


Samsung Adopts Iguazio’s Nuclio Serverless PaaS for Real-time Intelligent Applications Herzliya, March 7th 2019 – Iguazio, provider of the high performance platform for serverless and machine learning applications, announced it is partnering with Samsung SDS, a global software solutions and IT services company, to accelerate and streamline the delivery of intelligent applications. Samsung SDS has invested in Iguazio and will incorporate its platform into Samsung’s cloud services portfolio, powering serverless agility and data science operations for cloud native and AI-driven applications. "Samsung SDS is excited to invest in Iguazio. We look forward to providing our customers with intelligent, serverless applications by implementing Iguazio's technology to our cloud's PaaS," said Dr. Shim Yoon, Executive Vice President, Cloud Business Division Leader of Samsung SDS. “Iguazio welcomes Samsung SDS as a strategic Investor,” said Asaf Somekh, CEO, Iguazio. “We’re already working with different Samsung SDS groups on financial services and manufacturing deployments and are excited about the value Iguazio has created by powering Samsung’s cloud with serverless and machine learning. ” Iguazio’s platform includes data services and AI tools, empowering end-to-end serverless agility in the enterprise and real-time applications to improve performance, security, collaboration and the scalability of machine learning. Iguazio’s Nuclio is the leading open source serverless framework, enabling the development of modern applications over Kubernetes without having to manage infrastructure. About Samsung SDS Samsung SDS was founded in 1985 and has been leading the digital transformation and innovation of its clients for over 30 years. With the vision to become a...

---


Collaborating with Trax on a Kubernetes-powered hybrid cloud for real-time supply chain and intelligent operations KUBECON, Seattle, Tuesday, Dec 11, 2018 – Iguazio, the serverless platform for intelligent applications, today announced it is partnering with Google Cloud to enable real-time AI across the cloud and intelligent edge. Google Cloud and Iguazio’s hybrid cloud is enabling Trax, the leading provider of computer vision and analytics solutions for retail, to benefit from Kubernetes and a cloud-native architecture without managing its underlying infrastructure. Trax’s retail solutions leverage image recognition and predictive analytics to efficiently manage the physical shelf for consumer packaged goods manufacturers and retailers. “At Trax, we’re digitizing the world of retail by monitoring, predicting and optimizing store-and-field performance in real-time to improve on-shelf availability, optimize click-and-collect processes and modernize the shopping experience,” said Trax Chief Technology Officer, Yair Adato. “We recognized that we needed an edge-to-cloud solution that was built for speed, scale and intelligence – one that allowed us to focus on our application, versus the management of our infrastructure. ” “Kubernetes exemplifies the power of the consistent platform where customers appreciate learning once and use anywhere. Building applications on top of Kubernetes ensures companies can deploy workloads either on premises or in the cloud of their choice,” said Aparna Sinha, Group Product Manager, Kubernetes and GKE, Google Cloud. “The retail industry requires data portability across the cloud and intelligent edge. We are excited to collaborate with Iguazio to deliver a solution that enables real-time analytics of store data, all...

---


Leading open source serverless framework now includes capabilities that enable faster end-to-end enterprise and IoT deployments with reduced operational complexity SERVERLESS NYC, NEW YORK – October 30, 2018 – Iguazio, the Continuous Data Platform for real-time intelligent applications, today announced the release of a new version of Nuclio, the fastest fully integrated multi-cloud, on prem and edge enterprise serverless platform. Nuclio is already powering real-time intelligent operation management applications for financial institutions, telcos, transportation and manufacturing companies. Some examples include: Predicting and eliminating network latencies in electronic trading platforms to guarantee continuous trading. Predicting network failures, outages and cyber-attacks in robust telco operations, enabling proactive problem resolution. Real-time stockout predictions for intelligent supply chain applications. Predictive maintenance and corrective actions for real-time smart manufacturing. Nuclio gains enterprise capabilities when running on the Iguazio platform across all these use cases. It enables rapid development in the cloud alongside deployment at the edge for IoT, retail and manufacturing, delivering real-time analytics closer to the sources of data. “Nuclio is the first open, fast, enterprise-ready, ‘everywhere’ serverless framework,” said Iguazio’s CTO, Yaron Haviv. “This new version represents the second generation of serverless, taking usability, data integration and performance to the next level. We bring the serverless vision to enterprise applications and an end-to-end alternative over slower and cloud-specific solutions. Users seeking alternatives can now avoid tedious integration of point solutions. ” New features in this Nuclio release include: Security: authentication, authorization, data security, “dark-site” and offline deployment Persistent functions: with real-time access to...

---


Serverless NYC Speakers Include Google, IBM, Oracle and Capital One NEW YORK – October 10, 2018 – Iguazio, the Continuous Data Platform for real-time, intelligent applications, announced today Serverless NYC, a community-organized ServerlessDays conference taking place October 30th at Galvanize New York. Serverless NYC will reflect the state of serverless computing as it continues to generate buzz. It is a community-organized conference, attracting industry leaders such as Kelsey Hightower (Google), Jason Katzer (Capital One), Dave Grove (IBM) and Gwen Shapira (Confluent). Attending Serverless NYC are developers, architects, CTOs and engineers from a wide range of industries in the New York area. The conference will go beyond the hype to address real use cases ranging from success stories to challenges that have yet to be resolved. “Our goal is to make serverless available to the mainstream developer across multi-cloud and edge deployments,” said Yaron Haviv, Iguazio CTO. “We made this conference vendor neutral because serverless is still nascent. We’re bringing the community together to address issues currently holding serverless back from becoming the ubiquitous, mainstream development framework of the future. ” In addition to panel discussions about serverless implementations, the conference will include training sessions, during which developers can experiment with serverless frameworks on Nuclio, OpenWhisk, MongoDB Stitch, Spotinst and Microsoft Azure to expand their skillsets. “Serverless computing is the next step toward unshackling developers towards benefitting from the flexibility and elasticity of the cloud,” said Principal Ovum Analyst, Tony Baer. “While serverless may make infrastructure almost disappear, developers must understand...

---


Hybrid cloud solution enables leading Asian ride-hailing applications to deliver real time and event-driven insights SINGAPORE – August 15, 2018 – Equinix, Inc. (Nasdaq: EQIX), the global interconnection and data center company, today announced that it is collaborating with Iguazio to enable a single unified data platform to power continuous analytics and event-driven applications at scale. Iguazio’s Continuous Data Platform on Platform Equinix® enables a real time solution for high performance AI and event-driven applications to optimize performance in the smart mobility industry in Asia-Pacific. Ride-hailing services in Southeast Asia are expected to continue to grow in popularity, with six million rides booked through platforms daily, according to the “e-Conomy Southeast Asia Spotlight 2017” report by Google and Temasek. This, along with the booming internet economy in the region, is leading to an increased demand for interconnection that can enable businesses to securely and directly connect with customers, partners and service providers. Equinix Interconnection Oriented Architecture™ (IOA™) places strategic control points next to users, clouds and networks to build a digital edge. Iguazio, which simplifies the development and deployment of real time AI applications, leverages Platform Equinix to ensure that customers have the freedom to run at the edge for real time processing and reliability, while benefiting from multiple clouds for compute elasticity. This new hybrid cloud solution enables companies offering smart mobility to manage data intelligently and gain actionable business insights, ensuring reduced latency, maximized performance and fine-grained security. The combination of large data volumes and AI deployed in...

---


Nuclio enterprise edition gains momentum with leading cloud providers, as well as in on prem deployments and at the edge KUBECON AND CLOUDNATIVECON EUROPE, Copenhagen, Denmark – Wednesday, May 2, 2018 – Iguazio, the Continuous Data Platform for real-time applications, today released the enterprise version of Nuclio, its open-source serverless framework. Nuclio is the first fully integrated, cloud-neutral serverless framework for high-volume data processing, real-time analytics and artificial intelligence (AI). Iguazio and Microsoft have developed native integration for Nuclio with multiple Microsoft Azure services, providing the fastest, most advanced serverless framework deployable at the edge, in leading cloud vendors or on prem, while offering seamless workload portability among all three environments. Iguazio powers serverless for the enterprise with its Continuous Data Platform, simplifying operational challenges and enabling a faster delivery of solutions. “Nuclio is now running natively in the Azure Kubernetes Service and is fully integrated with Azure’s Application Insights,” said Liam Kelly, GM Commercial Software Engineering EMEA at Microsoft. “Iguazio is enabling Microsoft clients to rapidly build real-time applications, leveraging Azure’s native platform features such as events, monitoring and microservices. ” “Iguazio’s platform is unique in that it combines high performance and data services with serverless functions,” added Mika Borner, Management Consultant for Data Analytics at LC Systems. “We use Nuclio to develop solutions in the cloud for deployment either at the edge or on prem, allowing us to rapidly develop and operationalize applications. ” “We’re working to define standards for serverless with the Cloud Native Computing Foundation (CNCF),...

---


Sri Lanka’s highest performing mobility app uses iguazio to maintain its competitive edge and operational efficiency Herzliya, Israel – Wednesday, April 11, 2018 – iguazio, the provider of the Continuous Data Platform powering data-driven applications, today announced that it has provided its platform to PickMe, the leading on-demand transportation service in Sri Lanka. The PickMe ride-hailing app runs iguazio’s platform for data and operational efficiencies, including fraud detection and real-time heatmaps. The news of this partnership comes at a time when Asian internet connectivity is booming. As a result, the Southeast Asian market for connected vehicles has quadrupled in size since 2015 and will be valued at $20. 1 billion by 2025, according to a report co-authored by Google. At present six million rides are booked in Southeast Asia daily. PickMe’s deployment of iguazio enables the ride-hailing company to leverage iguazio’s real-time and unified database engine for rapid and actionable business insights via edge and hybrid cloud deployments. iguazio’s platform allows PickMe to combine historical and current data, using AI to detect and reduce fraud. iguazio’s smart mobility supply and demand heatmaps enable PickMe to maximize driver benefits and minimize passenger wait times. "As a fast growing business, it is important that PickMe obtain better and more efficient technologies that keep up with the pace of our growth,” said Jiffry Zulfer, CEO of PickMe. “By deploying iguazio we wish to do just that, resulting in seamless door-to-door transport so that our growth team can produce better results in real-time data...

---


Annual Guide Recognizes the IT Channel’s Top Partner Programs Herzliya, Israel, April 3, 2018 – iguazio, the provider of the Continuous Data Platform powering data-driven applications, announced today that CRN®, a brand of The Channel Company, has recognized iguazio in its 2018 Partner Program Guide. This annual guide is the definitive listing of partner programs from technology vendors that provide products and services through the IT channel. To compile the guide, The Channel Company’s research team assessed each vendor’s partner program based on investments in program offerings, partner profitability, partner training, education and support, marketing programs and resources, sales support and communication. iguazio’s channel partners are reselling its Continuous Data Platform globally and in a wide range of vertical markets such as financial services, IoT, manufacturing, telematics and cybersecurity. Its partner program is open to systems integrators, value-added resellers (VARs) and OEMs. “Finding the right technology vendor to partner with can be a daunting task, given the vast array of choices now available to solution providers,” said Robert Faletra, Executive Chairman of The Channel Company. “CRN’s Partner Program Guide helps them narrow the field and find the best fit, identifying the most rewarding partner programs and providing crucial insight into their strengths and benefits. We are pleased to present our 2018 PPG list, recognizing the strongest and most successful partner programs in the channel today. ” “We are delighted to be featured in CRN’s 2018 Partner Program Guide,” said Asaf Somekh, CEO of iguazio. “Our partners leverage iguazio’s technology, and...

---


Continuous Data Platform Now Available to More Enterprises Through Systems Integrators, VARs, OEMs HERZLIYA, February 28th, 2017 – iguazio, the leading data platform powering data-driven applications by extending the cloud experience at the edge and on premises, today unveiled its global Channel Partner Program. As part of the new program, which is open to systems integrators, value-added resellers (VARs) and OEMs, iguazio has already formed partnerships with key channel leaders. iguazio’s channel partners are reselling its Continuous Data Platform globally and in a wide range of vertical markets such as financial services, IoT, manufacturing, telematics and cybersecurity. The iguazio platform accelerates the deployment of machine learning and artificial intelligence, eliminating data pipeline complexities and reducing time to insights. The Continuous Data Platform is a turnkey solution that enables customers to ingest, enrich and analyze streams, objects, files and database records - all in one simple, fast and secure platform. “Our channel program’s regional and vertical focus addresses the needs of iguazio’s early customers and prospects in the US, Europe and Asia,” said Asaf Somekh, CEO of iguazio. “Our partners leverage iguazio’s technology, and we’ve been thrilled to see them expand upon it by creating value and building integrated solutions for robust, distributed data architectures in dedicated industries, such as manufacturing and financial services. ” With its Channel Partner Program, iguazio provides hands-on engagement, training, certification and support to ensure partner success while seeking and building additional strategic relationships. The iguazio’s iCDE certification (Certified Data Engineer) ensures iguazio’s partners and customers...

---


Demand for iguazio’s Hybrid Cloud and Edge Solutions Drives Continued Global Expansion STRATA DATA CONFERENCE, Singapore, December 6th, 2017 – iguazio, the leading data platform powering continuous analytics and event-driven applications, announced today the opening of its Asia Pacific (APAC) regional headquarters in Singapore. iguazio announced Singapore-based Grab as a marquee customer in September 2017 and is rapidly growing its customers across industries including financial services, automotive, IoT and telecommunications that require real-time data processing, artificial intelligence and machine learning. In July iguazio raised US $33 million in Series B funding from investors including Verizon Ventures , Robert Bosch Venture Capital GmbH, the Chicago Mercantile Exchange Group, Dell Technologies Capital and leading venture capitals, bringing total funding to US $48 million. Funding from the latest round is being used for the company’s global expansion. “We’re seeing great demand in the APAC region, which will generate a third of global revenues related to data driven applications,” said iguazio CEO Asaf Somekh. “Our APAC expansion better positions iguazio to work with innovative companies such as Grab which are driven by the growing up-take of digital technologies among consumers and businesses. We’ve chosen Singapore as our regional headquarters due to its role as a hub for innovation, opening up a variety of opportunities including cybersecurity, IoT, Smart City, financial services and telematics, making it an ideal location to accelerate iguazio’s APAC operations. ” Grab is one of several customers within a variety of industries that iguazio’s platform is enabling. iguazio provides Grab with...

---


nuclio expands iguazio’s data platform to provide a complete cloud experience that allows faster, flexible deployment in the cloud, on prem or at the edge KUBECON AND CLOUDNATIVECON, Austin, December 6, 2017 – iguazio, the leading data platform enabling continuous analytics and event driven-applications, today unveiled nuclio: its open source, ultra-fast, multi-cloud serverless platform. nuclio functions are faster than bare-metal, processing up to 400,000 events per second with a single process, address a broad set of applications, are simpler to develop and can be deployed anywhere. Building modern applications with nuclio minimizes operational overhead and frees developers from lock-ins to cloud specific APIs or services. iguazio’s data platform converges and simplifies data management. With nuclio, iguazio completes a full-blown cloud experience of data services, AI and serverless - all delivered in one integrated and self-managed offering, at the edge or in a hosted cloud. “Customers tell us they want to bring the cloud experience to their data, not force fit all their workloads into the public cloud,” said SiliconANGLE Media CEO Dave Vellante. “In our opinion, iguazio’s data and serverless platforms are a leading example of just that; a set of on-prem services that enable modern application delivery with a true cloud experience close to data sources. ” nuclio’s top strengths include: A real-time function OS delivering 30-100x faster performance and lower latency when compared to other public cloud or FaaS offerings (see benchmark) Automated function deployment, scaling and operations Support for a large variety of open or cloud-specific event...

---


Early customer adoption includes Grab, the Largest Ride-Hailing Service in Southeast Asia STRATA DATA CONFERENCE, New York, September 27, 2017 – iguazio, the leading data platform for continuous analytics and event driven-applications, announced today the general availability of its Unified Data Platform. iguazio simplifies data pipeline complexities while providing a turnkey solution to accelerate the development and deployment of machine learning and artificial intelligence in the enterprise, generating fresh insights in real-time. In a separate press release, iguazio today also announced Grab, the Singapore-based ride-hailing giant, has selected iguazio’s Unified Data Platform to simplify and accelerate its data pipeline and analytics. See related press release here. “We have yet to see anything that compares directly to iguazio’s combination of data analytics and cloud architecture,” said Matt Aslett, Research Director, Data Platforms and Analytics at 451 Research. “iguazio’s Unified Data Platform is differentiated and focused on reducing the complexity involved in deploying multiple data processing technologies, while also addressing machine learning and data science, so that customers can act on insights instead of spending time and resources on data management. ” The iguazio Unified Data Platform: The iguazio Unified Data Platform comes with fully integrated essential applications dynamically deployed over Kubernetes, including artificial intelligence and machine learning frameworks like Spark, R and TensorFlow, visualization and a real-time serverless framework. iguazio’s solution enables the entire continuum of analytics - including predictive and prescriptive analytics - through the ingestion and enrichment of streams, objects, files and database records. Previewed at the Strata Data...

---


Largest Ride-Hailing Service in Southeast Asia Uses iguazio’s Platform to Ingest, Enrich and Analyze Data for Continuous Analytics and Event-Driven Applications STRATA DATA CONFERENCE, New York, September 27, 2017 – iguazio announced today that Grab, the leading on-demand transportation and mobile payments service in Southeast Asia, has selected iguazio’s Unified Data Platform to accelerate innovation and boost its competitive edge in a market that serves more than 600 million consumers. The iguazio Unified Data Platform deployed at Grab – which facilitates more than 3 million rides per day -- is being used for a variety of data monetization and operational efficiency objectives. “Ride hailing is more than just an app – there’s a lot going on in the background that leads to great experiences for both the passenger and driver,” said Ditesh Gathani, Director of Engineering at Grab. “Grab’s business thrives because of our ability to manage a continuous stream of data that facilitates real-time matching, booking, payment and location services. iguazio will enable us to continue offering new innovative services at a faster time to market. ” Grab selects iguazio By selecting iguazio’s Unified Data Platform, Grab is innovating a variety of applications, such as: Driver incentives to determine driver bonuses in real-time, by analyzing and increasing driver effectiveness, the number of runs during peak driving times and rider satisfaction. Maximize driver profits while reducing passenger wait times by optimizing driver the decision-making process using advanced real-time supply and demand heatmaps. Surge pricing optimization which correlates passenger demand data...

---


New round includes strategic investors from financial services, IoT and service providers following successful early deployments HERZLIYA, July 25th, 2017 iguazio, a global pioneer in real-time edge analytics, announced today a Series B investment of $33 million led by Pitango Venture Capital, with additional funds from Verizon Ventures, Robert Bosch Venture Capital GmbH (RBVC), CME Ventures and the company’s existing investors, Magma Venture Partners, Jerusalem Venture Partners and Dell Technologies Capital. iguazio accelerates the digital transformation of enterprise companies and simplifies real-time analytics at the edge, on-premises and in hybrid environments, complementing the offering of leading cloud providers. This new financing brings the company’s total investment to $48 million. “iguazio’s team has an outstanding track record of innovation and execution and we are delighted to back these stellar managers once again,” said Eyal Niv, Managing General Partner at Pitango. “While the majority of big data deployments fail due to over complexity, iguazio’s platform has proven to be simple, fast and secure, making it exceptional for artificial intelligence and machine learning use cases. We’ve already received overwhelming feedback from beta customers generating actionable real-time insights with significant business impact. ” “As one of the largest telecom companies in the world, we witness the importance of real-time continuous analytics and the way it has become crucial across businesses. Yet, there are not many existing scalable solutions,” said Merav Rotem-Naaman, Managing Director at Verizon Ventures Israel. “iguazio is aiming to become a trusted partner for companies looking to use data to make real-time...

---


Delivering continuous analytics and offering greater simplicity, performance, security and agility for next generation applications HERZLIYA, May 4, 2017 iguazio today announced it has been included in the “Cool Vendors in Data Management, 2017” report by Gartner, Inc. The company’s real-time continuous analytics platform simplifies the data pipeline to accelerate business insights and enhance security. According to the Gartner Cool Vendors in Data Management, 2017 report, “unified data access, multimodel DBMS approaches, metadata management techniques and real-time data integration top the list of innovations for CIOs and data and analytics leaders. ” The report also finds that “real-time data integration and access to data remain core challenges for data and analytics leaders looking to modernize data management ecosystems. ” “We are honored to be recognized by Gartner as a Cool Vendor in Data Management,” said Asaf Somekh, CEO of iguazio. “We’ve been helping our customers in IoT and financial services to dramatically simplify their data pipeline and generate real-time business insights. iguazio has taken the bold approach of completely redesigning the data stack with our high-performance unified data model architecture. Today’s digital transformation requires agile data consumption and we believe being one of Gartner’s Cool Vendors in Data Management is a testament of iguazio’s disruptive technology. ” Data is at the center of iguazio’s continuous analytics platform: iguazio ingests, enriches, analyzes and serves data – securing data and allowing access to the same records simultaneously through streaming, object, file and database APIs in real-time. The iguazio solution integrates with the...

---


Complete re-thinking of the traditional data pipeline reduces time-to-insights from hours to seconds STRATA+HADOOP WORLD, SAN JOSE, March 14, 2017 iguazio is demonstrating its groundbreaking continuous analytics solution at Strata + Hadoop World. Designed to help enterprises solve big data operational challenges and generate real-time insights, iguazio’s real-time continuous analytics platform reduces time to insights from hours to seconds, eliminating data pipeline complexities, while seamlessly integrating with Apache Spark and Kubernetes. Data is at the center of iguazio’s real-time continuous analytics platform, which simplifies the data pipeline and speeds it up: iguazio ingests, enriches, analyzes and serves data – all in one unified platform. It integrates with the open-source frameworks of Spark and Kubernetes to accelerate insight generation and enable rapid deployment of a variety of stateless analytics services and data processing tasks. iguazio’s platform secures the data and allows accessing the same records simultaneously through streaming, object and database APIs. “Digital transformation is leading us to agile data consumption, and continuous analytics is its cornerstone. In different industries, especially in financial services, healthcare and IoT, organizations have a need to tackle the challenge of complexity across the entire data lifecycle,” said John L. Myers, Managing Research Director at Enterprise Management Associates – a Boulder, CO based industry analysis firm. “iguazio’s continuous data consumption approach empowers organizations to ingest data into a unified repository and run multiple stateless processing engines which enrich, aggregate, infer and act on the data to enable this transformation. ” iguazio early deployment customers benefitting from...

---


Delivers 100x faster performance and 10x lower cost for on-premises and hybrid cloud deployments STRATA+HADOOP WORLD NEW YORK 2016, NEW YORK, September 27, 2016 iguazio today introduced its flagship product, the Enterprise Data Cloud platform, unleashing the full potential of megatrend applications and analytics for big data, the Internet of Things (IoT) and cloud-native applications. iguazio has pioneered a new service-driven approach to enterprise data management, redesigning the entire data stack to accelerate performance and bridge the enterprise skill gap. iguazio’s Enterprise Data Cloud is the only secure data platform-as-a-service deployed either on-premises or in hybrid cloud architectures, with self-service portals and APIs for developers and operators. Previewed at Strata + Hadoop World New York 2016, the Enterprise Data Cloud represents a shift in the way enterprises are storing and analyzing data, where information often is stored in silos, is difficult to access and is expensive to maintain. The new unified platform delivers a breakthrough in application performance and efficiency. With only four data appliances, companies can store up to 10 petabytes per rack, with costs starting at $0. 03 per gigabyte per month. The platform delivers 10 million transactions per second and throughput of 50 gigabytes per second with sub 100 microsecond application latencies, across streaming, NoSQL, objects or files. iguazio’s revolutionary stack, already in early customer deployments, supports simultaneous high-performance access through multiple industry standard and Amazon-compatible APIs. “With iguazio’s Enterprise Data Cloud, our enterprise customers have the ability to build large data repositories closer to data sources...

---


Reveals Details of Extremely Efficient Architecture that Seamlessly Accelerates Spark and Hadoop, Busts Silos and Ends ETL SPARK SUMMIT 2016, SAN FRANCISCO, June 7, 2016 iguaz. io, the disruptive company challenging the status quo for big data, the Internet of Things (IoT) and cloud-native applications, today unveiled its vision and architecture for revolutionizing data services for both private and public clouds. This new architecture makes data services and big data tools consumable for mainstream enterprises that have been unable to harness them because of their complexity and internal IT skills gaps. Data today is stored and moved between data silos optimized for specific applications or access patterns. The results include complex and difficult-to-maintain data lakes, constant data movement, redundant copies, the burdens of ETL (extract/transform/load), and ineffective security. While popular cloud services like Amazon’s AWS and Microsoft’s Azure Data Lake introduce some level of simplicity and elasticity, under the hood they still move data between different data stores, lock customers in through proprietary APIs and onerous pricing schemes, and, at times, provide unpredictable performance. Data is proliferating at an unprecedented pace — analyst firm Wikibon predicts the big data market will grow to $92. 2b by 2026 — requiring a new paradigm for building and managing a growing and complex environment. With its first-ever high-performance virtualized data-services architecture, iguaz. io is taking a fresh approach to big data’s challenges. The iguaz. io architecture: Consolidates data into a high-volume, real-time data repository that virtualizes and presents it as streams, messages, files,...

---


Iguaz. Io raises $15 million in series a funding to disrupt big data storage Herzliya, Israel. Nov. 25, 2015 – Iguaz. io, a provider of innovative data management and storage solutions for Big Data, IoT and cloud applications today announced a $15 million Series A funding round. Led by Magma Venture Partners, the funding includes additional investments from JVP and large strategic investors. The iguaz. io founding team is comprised of a group of former executives from successful technology companies in the fields of storage, cloud computing, high-speed networking, analytics and cyber-security. These companies include XtremIO (acquired by EMC), XIV (acquired by IBM), Mellanox (NASDAQ: MLNX), Voltaire (NASDAQ: VOLT, acquired by Mellanox) and Radvision (acquired by Avaya). Over the past two decades, the team has enabled enterprise customers to evolve their data centers and has helped leading cloud operators design and deploy their hyper-scale data centers. Iguaz. io will use the funding to continue growing its team of experts in the spaces of Big Data, storage, security and networking (see www. iguaz. io/careers). According to IDC, the market for Big Data will surpass $41 billion by 2018, growing six times faster than the overall IT market. “Big Data and cloud computing are creating tectonic shifts in the market, setting the stage for market disruptions,” said Yahal Zilka, Managing Partner at Magma Venture Partners. “The iguaz. io team, with its innovative approach, is well positioned to disrupt the market. ” “Iguaz. io has attracted top talent with diverse, multidisciplinary skills and...

---


---


## Sessions


---


## MLOps Terminologies


", "_top_content": "field_6019145ebd0a7", "list_items": "", "_list_items": "field_6019148cbd0a9", "bottom_content": "", "_bottom_content": "field_60191482bd0a8", "margin_top": "mt-default", "_margin_top": "field_5e3c7bd8719f4", "margin_bottom": "mb-default", "_margin_bottom": "field_5e3c7c33719f5", "padding_top": "pt-default", "_padding_top": "field_5e3c7c99719f6", "padding_bottom": "pb-default", "_padding_bottom": "field_5e3c7ce5719f7", "block_id_block_id": "", "_block_id_block_id": "field_62a1c4463df6f", "block_id": "", "_block_id": "field_62a2f514c3ecb" }, "align": "", "mode": "auto" } /-->

---


", "_top_content": "field_6019145ebd0a7", "list_items": "", "_list_items": "field_6019148cbd0a9", "bottom_content": "", "_bottom_content": "field_60191482bd0a8", "margin_top": "mt-default", "_margin_top": "field_5e3c7bd8719f4", "margin_bottom": "mb-default", "_margin_bottom": "field_5e3c7c33719f5", "padding_top": "pt-default", "_padding_top": "field_5e3c7c99719f6", "padding_bottom": "pb-default", "_padding_bottom": "field_5e3c7ce5719f7", "block_id_block_id": "", "_block_id_block_id": "field_62a1c4463df6f", "block_id": "", "_block_id": "field_62a2f514c3ecb" }, "align": "", "mode": "auto" } /-->

---


hbspt. cta. load(5272961, '20436803-fd1f-4c93-906c-39ccd3b0a5e1', {\"useNewLoader\":\"true\",\"region\":\"na1\"}); \r\nFeature Vectors in Exploratory Data Analysis\r\nIn exploratory data analysis, researchers try to discover features from raw data. They may start with qualitative research, looking at visualizations and applying their domain expertise to deduce an idea that can transform the observation into feature vectors. For example, a feature vector in data mining represents a hidden pattern in large data sets, such as equity trading buy\/sell signals from the historical trading price and volume data. \r\n\r\nIn the field of natural language processing, the process of splitting sentences into distinct entities is known as tokenization. For instance, researchers could treat each word or phoneme as a unique token to generate feature vectors for further analysis and experiments. \r\n\r\nIn computer vision, the RGB color scheme isn’t the only way to represent image pixels. For example, there is also HSL (hue, saturation, lightness) and HSV (hue, saturation, value). Sometimes, practitioners even use a monochrome scheme to reduce noises originating from color images. \r\n\r\nUltimately, researchers explore different feature vectors to evaluate the performance of their predictive models. Once the feature design is ready, they are good to go to the next stage. \r\nFeature Vectors in Feature Engineering\r\nFeature engineering is, in large part, the systematic process of generating feature vectors from raw data. There are, however, some obstacles to setting up such a process. First, we need a place to store generated feature vectors for later retrieval. We also need to update feature definitions from time to time to accommodate changes in...

---


Want to learn more about real-time machine learning? Book a live demo here.

---


Want to learn more about industrializing unsupervised machine learning models? Book a live demo here.

---


Want to learn more about extending Kubeflow into a full MLOps platform? Book a live demo here.

---


---


## Case Studies


---


## Solutions


---


## Q&A


Traditional machine learning (ML) has been reliably used for computational tasks due to its deterministic and reproducible nature. These models are excellent at handling numbers and structured outputs, which can then be leveraged by LLMs to ground responses in factual data, preventing LLM hallucinations. The key approach is to use the best tool for the job—traditional ML for structured, explainable tasks and LLMs for generative tasks. Explainability is crucial, as traditional ML models are much easier to interpret than LLMs, making them valuable in contexts where transparency is required. For more on explainability, see here. For a deeper dive on combining traditional ML with LLMs, check out this blog. How do gen AI and traditional AI complement each other? Check out this related question.

---


When designing a multi-agent system for process automation, such as in a contact center, the process typically starts with analyzing historical data (e. g. , call transcripts, user interactions) to understand how human agents complete tasks. An LLM-based model is often used to analyze this data and identify the common steps a person takes. For example, looking at calls from the past six months and understanding how agents authenticate users, determine intent, etc. Since there is variability in execution, a probabilistic model helps map out these actions in a structured way. Once a granular workflow is defined, SMEs validate and refine it, creating a detailed blueprint of human activities. This blueprint is then used to design agent milestones. First, deciding whether workflows require one agent or a sequence of agents. If the workflow includes multiple tasks, choose an architecture with a sequence of agents. Each agent should be assigned a specific task that can perform well to optimize efficiency—e. g. , one for user authentication, another for sentiment analysis, and another for breaking down complex requests. Finally, an orchestrator integrates these agents into a seamless system.

---


Open-source MLRun can be used for efficient resource management in a number of ways. A few examples include: Auto-scaling - Automated resource allocation based on workload needs. Experiment tracking to compare models and choose the best-performing one without re-running the entire training pipeline. Serverless deployments with auto-scaling. Support for model quantization and pruning. Monitoring and logging for resource usage. Parallel pipeline execution and distributed compute capabilities. Micro-batching - Processing multiple requests simultaneously, improving GPU utilization and lowering per-request costs. Read more about auto-scaling GPUs, experiment tracking, and how to use open-source Nuclio for serverless deployment.

---


Unlike traditional machine learning, generative AI introduces unique challenges and opportunities, necessitating specialized skills and a proactive approach to learning. To upskill talent for generative AI, training recommendations vary by role: Data scientists working with generative AI must go beyond standard analytical methods, developing expertise in prompt engineering and model output monitoring. This might require adapting AI model monitoring practices to ones that are specific to generative AI. For example, addressing issues like hallucinations. Data engineers, software engineers and cloud architects focus on managing complex infrastructure needs for AI deployments. This requires staying current with cloud service offerings, AI-specific pipelines and model testing is important, as these are rapidly evolving. This knowledge is essential for transitioning from experimental models to production-ready deployments. Given the rapid pace of advancement in the AI field, it’s critical for all professionals working with generative AI to prioritize continuous learning. This can be achieved by regularly engaging with content from industry analysts, AI vendors and thought leaders in the space. Blogs, newsletters, webinars and communities provide real-time insights into best practices, new tools and emerging trends. Additionally, attending industry events and joining online communities, such as AI research forums or professional groups on LinkedIn, helps professionals network with peers, share insights, and exchange knowledge about challenges and solutions.

---


Gen AI positively impacts operational and business performance. For example, companies using gen AI-powered operator and technician co-pilots in maintenance have seen significant productivity and efficiency gains. These AI tools combine data from manuals, operating procedures and past root cause analyses, suggesting hypotheses and scenarios for troubleshooting. This allows technicians to troubleshoot issues more quickly and effectively on the production line. In traditional troubleshooting scenarios, root cause analyses could take hours or even days, but with Gen AI assistance, these steps are streamlined to seconds. This reduces breakdown time and boosts productivity, with some companies reporting up to 30-40% improvements in troubleshooting speed and 5-10% increases in line output due to faster response times. Similar gains have been observed in contract feedback and improvement processes, further enhancing productivity across value chains. Although these are still individual use cases, broader end-to-end transformations integrating these capabilities could yield even greater impacts in the future.

---


Gen AI and traditional AI serve different purposes. They can be used separately and together. Traditional AI, such as classification and ML models, excels at specific tasks like pricing optimization and customer segmentation. These models are designed to be precise, stable and reliable for data-driven tasks. Gen AI, on the other hand, can be used for efforts like content creation, chatbots and agents, as a virtual assistant and more. Gen AI complements traditional AI by providing interpretive capabilities. For example: Generating informative reports or summaries based on traditional AI results A customer-facing chatbot that can route customer requests to traditional AI models tand then answer queries. For example, about package shipping times Generating synthetic data Analysis of the sentiment behind customer reviews, feedback, social media mentions, etc. Automate the labeling of textual data with high accuracy. See how this works here. Rather than replacing traditional AI, Gen AI enhances its output, making insights more accessible and adding flexibility to existing systems.

---


Gen AI outputs need to be evaluated for accuracy, relevancy, comprehensiveness, how they de-risk bias and toxicity, and more. This should be done before they are deployed to production and acted on, to avoid issues like performance, ethical matters, legal issues and disruptions. The methods that can be used to evaluate outputs include: Comparing the results to the data source they were retrieved from Ensuring consistent responses by running similar prompts multiple times Using LLM-as-a-Judge to allow an additional LLM to evaluate the results Testing outputs and fine-tuning the model for adherence to industry-specific knowledge requirements or specific brand voice Reviewing responses against a checklist of essential components for the given topic or field Implementing guardrails and filters for toxicity, hallucinations, harmful content, bias, etc. Implementing guardrails for security and privacy Continuous monitoring and feedback loops to ensure ongoing quality and relevancy Establishing LLM metrics to track the overall success of the AI model in meeting its intended purpose

---


Effective gen AI guardrails are required throughout the data, development, deployment and monitoring AI pipelines These guardrails help in mitigating risks such as biased outputs, data privacy breaches, and toxic content generation. Here are some key examples of guardrails that can be implemented in Gen AI pipelines: Prompt Engineering - Crafting input prompts to guide the AI model toward generating desirable outputs. These prompts discourage the model from generating toxic or biased outputs and provide context that steers the model in a safer direction. LLM as a Judge - Using an LLM to evaluate the output’s compliance with predefined rules and standards. When a violation is detected, the LLM flags the issue or provides recommendations for modification to ensure alignment with the guidelines. Toxicity Measurement with Language Filters - Implementing language filters to measure and flag toxicity levels in generated outputs. These filters use pre-trained models or rule-based systems to detect offensive, harmful, or otherwise inappropriate language. Bias Detection and Mitigation - Techniques such as adversarial testing, fairness metrics and model retraining to identify and mitigate biases within AI models. This involves detecting biases towards specific populations, genders, or other demographic groups. Data Privacy Checks - Checks for sensitive information. This could involve using regular expressions (regex) to identify and redact personal information such as social security numbers, addresses, and other private data. This step allows complying with regulations like GDPR or CCPA and protecting user privacy. Hallucination Detection - Using knowledge bases or fact-checking algorithms to compare generated content...

---


When deploying LLMs in production, monitoring prevents risks such as malfunction, bias, toxic language generation, or hallucinations. This allows for tracking application performance and helps identify issues before they impact users. In addition, logging interactions allows understanding user queries and the responses, which can be used for fine-tuning the model. There are multiple levels of monitoring: Functional Monitoring - Ensures that the LLM is operating correctly and identifies what models and versions are currently in use within applications. This helps manage risks if a model version becomes problematic or poses a security threat. Governance and Compliance - Centralized governance, according to the "AI Factory" approach. This helps organizations know which models are being used, which versions, when updates or patches are needed, or when security risks require patching and action. Resource Monitoring - Involves tracking the consumption of resources like CPU and memory by different applications or departments. This helps in identifying inefficient use of resources or applications that might be consuming too much without adding sufficient value.

---


In RAG, the model queries an external dataset or knowledge base, typically using a vector space model where documents or data snippets are encoded as vectors. When a query (prompt) comes in, the model performs a similarity search to retrieve the most relevant information. This retrieved content is then used by the generative component of the model to craft a response that is informed by this external data. RAG is particularly useful in scenarios where the required knowledge is vast or frequently updated, such as news updates, scientific research, or detailed customer data. It ensures that the model can provide accurate and relevant information without needing frequent retraining. Fine-tuning is a technique where a pre-trained model is further trained (fine-tuned) on a smaller, specialized dataset. This secondary training phase adjusts the model’s weights to perform better on tasks specific to the characteristics of the new data. Fine-tuning involves continuing the training process of an already trained model but focuses on a narrower scope or a specific domain. This targeted training helps the model to better understand and generate responses that are aligned with the specific nuances and requirements of the target domain. Fine-tuning is helpful when the model needs to adopt a specific tone, style, or set of knowledge, such as legal terminology, technical support for a specific product, or company-specific guidelines. Fine-tuning is also used to implement "guardrails" or constraints that guide the model’s outputs to avoid undesirable content or biases. RAG and retraining are not mutually exclusive and...

---


Start by evaluating the LLM and understanding how well it performs. This involves testing the model with various inputs to understand its strengths and weaknesses. Establish clear metrics that reflect the goals you aim to achieve with the LLM. For instance, if the LLM is used for customer care, metrics could focus on accuracy and response time. Break the overall problem into small tasks. This approach allows you to address specific areas where the LLM underperforms. Optimize the LLM with different prompts. Experiment with different prompting strategies to find the most effective ways to guide the model’s responses. This might involve tweaking the length, specificity, or format of prompts. You can also consider innovative approaches like fine-tuning the model on specific datasets, incorporating external knowledge bases, or using ensemble techniques where multiple models or prompts are used to generate a single output. Analyzing the model's responses to these variations can provide insights into more subtle aspects of its behavior and how it interprets instructions. Implement feedback loops where the model’s predictions are regularly reviewed and corrected by human operators. This real-world feedback can be used to further train and refine the model. Looking for tools to assist with LLM prediction accuracy? Try open-source DeepEval or RAGAS. Additional metrics that can help you evaluate your LLM include Answer Relevancy, Faithfulness, Hallucination, and Toxicity. (These are all covered in DeepEval). These metrics can be especially helpful when dealing with unstructured text dat. Unstructured data it requires something to compare against. However, a...

---


Call center companies can use gen AI for a wide range of use cases. Gen AI can assist with demand reduction, deflection and routing of calls, vendor optimization, improving operational efficiency and service optimization. Some of the most popular applications include: Co-pilots - Providing assistance and suggestions to agents in real-time based on customer history, product capabilities, contracts, policies, etc. Post-call insights - Taking the data from conversation, emails and chats, embedding into business analytics and deriving insights. Chatbots - Augmenting the use of automated chatbots to answer customer requests. Training - Simulations for new and experiences agents Companies that are able to combine their domain expertise with an understanding of their pain points, the gen AI tech and the future roadmap of gen AI products and the opportunities they hold, will be able to derive business value from gen AI apps.

---


Gen AI has opened up new opportunities for segmentation and personalization. Access to real-time behavioral data and analytics allows catering to users based on what is appropriate for them at this given moment. Preferences can change for the same user during different moments, e. g preferring text at one moment and a call at another. With hyper-personalization, organizations can provide an accurate and high-quality user experience, building trust, enhancing customer loyalty and growing their revenue. Hyper-personalization goes beyond the personalization techniques of the previous wave of AI, by understanding and catering to the nuances of individual preferences at a much higher scale. It represents a shift towards more intuitive, user-centric models of interaction, where services (like customer call centers) and content evolve along with the needs and desires of each user, enhancing user engagement, satisfaction, and loyalty. Using the insights gained from predictive modeling, generative AI systems can create or modify content, products, or services in real-time to match the individual needs of each user (to see an example of this, watch this demo of a call center agent co-pilot).    Hyper-personalization systems are dynamic, continuously learning from new data and user interactions. This allows the generative AI to refine its predictions and customizations over time, improving the relevance and accuracy of the personalized experiences it provides.

---


There are several risk factors to avoid when implementing gen AI. The most important ones are accuracy and hallucination. Despite the technological advancements, models hallucinate extensively and there are currently no good solutions for addressing this. Solutions like LLM-as-a-judge or RAG are incomplete, and many organizations put humans in the loop to label and tweak answers. This can also help with solving issues like data privacy, toxicity and bias. On the business level, enterprises must be aware that gen AI comes with unique risks that require a comprehensive strategic approach. Some of the most critical strategic initiatives are: Data governance: Implement robust data governance practices to ensure the quality, integrity, and security of data used by Gen AI systems. This includes data collection, storage, processing, access control, and compliance with data protection regulations. Security measures: Implement robust security measures to protect Gen AI systems from cyber threats, data breaches, unauthorized access, and malicious attacks. This includes encryption, access controls, authentication mechanisms, secure coding practices, and regular security audits. Continuous monitoring and evaluation: Continuously monitor and evaluate the performance, effectiveness, and impact of Gen AI systems in enterprise applications. Identify emerging risks, trends, and issues and take proactive measures to address them. For an overview and a few examples of gen AI risks, read the blog Implementing Gen AI in Practice For a detailed approach, check out the book “Implementing MLOps in the Enterprise” by Yaron Haviv and Noah Gift.

---


Iguazio supports any and all languages for gen AI applications. Today, there are projects in English, Turkish, Arabic, Portuguese, Spanish, Hebrew, Norwegian and more. New languages are implemented either through translations or by fine-tuning models in the new language. Fine-tuning and implementing LLMs in non-English languages involves several key steps and considerations: 1. Fine-Tuning: Fine-tune the selected LLM on the collected data in the target language. This involves training the model on language-specific tasks or objectives, such as language modeling, text classification, machine translation, etc. 2. Evaluation: Evaluate the fine-tuned model on appropriate benchmarks or validation datasets to assess its performance, accuracy, and generalization ability in the target language. 3. Language-specific challenges: Address any language-specific challenges or nuances during the development process. These may include morphological complexity, syntactic differences, lack of labeled data, and domain-specific terminology. 4. Adaptation: Adapt the model to specific applications or use cases in the target language. This may involve domain adaptation, transfer learning, or customization of model outputs to meet the requirements of the application. 5. Testing and iteration: Test the developed LLM application rigorously in real-world scenarios to identify and address any issues or limitations. Iterate on the development process as needed to improve performance and user experience.

---


The acceptable level of PII reduction accuracy in an AI pipeline depends on various factors, including the specific use case, industry standards, regulatory requirements, and the sensitivity of the data involved. Generally, a higher level of accuracy is desirable to minimize the risk of exposure of sensitive information. However, it's essential to balance accuracy with other considerations such as performance, efficiency, and usability. Here are some factors to consider when determining acceptable PII reduction accuracy: 1. Regulatory requirements and industry standards: If your AI pipeline deals with PII, it must comply with relevant data protection regulations such as GDPR, CCPA, HIPAA, etc. These regulations may specify certain standards or requirements for the handling and protection of PII. Some industries may have established best practices or standards for handling sensitive data, including PII. Compliance with these regulations and standards may therefore necessitate a higher level of accuracy in PII reduction. 2. Risk assessment: It’s a good practice to conduct a risk assessment to evaluate the potential consequences of inaccuracies in PII reduction. Consider the impact on individuals' privacy, the organization's reputation, legal liabilities, and the likelihood of data breaches or misuse. 4. Data sensitivity: The sensitivity of the PII involved should also influence the acceptable level of accuracy. Highly sensitive information, such as financial data or health records, may require a higher degree of accuracy compared to less sensitive information like public contact details. 5. Performance trade-offs: Achieving higher accuracy in PII reduction may come with trade-offs in terms of computational...

---


A Large Language Model (LLM) pricing strategy refers to how a company or organization determines the cost associated with using or accessing their language model, such as GPT-3 or similar models. Pricing strategies for LLMs can vary depending on factors such as the type of usage, the level of access, the volume of usage, and the target customer segments. Here are some common LLM pricing strategies: Pay-as-You-Go: In this pricing model, users are charged based on their actual usage of the LLM. It may involve pricing per token generated, per API request, or per hour of usage. This approach is flexible and allows users to pay only for what they use. Subscription Model: Some LLM providers offer subscription plans where users pay a fixed monthly or annual fee to access the model with certain usage limits. Subscriptions can offer cost predictability and may come with tiered pricing based on usage levels. Tiered Pricing: LLM providers may offer multiple pricing tiers with different features and usage limits. Customers can choose the tier that best suits their needs and budget, with higher tiers typically offering more features and higher usage limits at a higher price. Freemium Model: Some providers offer a free tier with limited functionality or usage to attract users, while offering premium paid tiers with more advanced features or higher usage limits. This strategy aims to convert free users into paying customers. Enterprise Plans: LLM providers may offer customized pricing for large enterprises or organizations that require high-volume access, specialized...

---


Open source Large Language Models (LLMs), like any other software, can pose security risks if not properly managed and used. Here are some of the potential security risks associated with open source LLMs: 1. Malicious use of generated content: Open source LLMs can be used to generate fake news, phishing emails, spam, or other malicious content. This can be a security risk as it can deceive users and spread misinformation. 2. Bias and discrimination: LLMs trained on open source data may inherit biases present in the training data, leading to biased or discriminatory outputs. This can have ethical and legal implications and potentially harm individuals or groups. 3. Privacy concerns: If an open source LLM is used to generate text that includes personal or sensitive information, there can be privacy concerns if this information is not properly protected or handled.   4. Unauthorized access: If an open source LLM is deployed on a server or in a cloud environment, there is a risk of unauthorized access to the model or the data it processes. Security measures must be in place to prevent such breaches. 5. Model poisoning: Malicious actors can attempt to manipulate the training data or the model itself to inject malicious code or biases. This can result in the generation of harmful or malicious content. 6. Data leaks: Open source LLMs may inadvertently leak sensitive information or proprietary data if not properly configured and secured. This could occur through generated text or via attacks on the model itself....

---


User intent is a classification problem. By giving the LLM model different examples of classifications, it can help simplify the process, since LLMs understand nuances. For example, by classifying the difference between booking a ticket and booking a restaurant, LLMs can understand that “booking a table” means “booking a restaurant”, even though there is no use of the word “restaurant” in the statement. For more information, NLP has a vast amount of literature dedicated to determining the most efficient classification technique to recognize user intent from a sentence.

---


Go to the vector search Gather three to five chunks from the vector search. Feed them into the LLM for realization Ensure a proper response. If you weren’t able to select the correct chunks, you can: Extract information from the question to refine the search Apply keywords in the indexing For example, let’s assume you have Uber’s financial reports for 2021 and 2022. If a user asks a question about Uber’s financials in 2022, you can extract “2022” and ‘Uber” from the question and use them as a keyword search to filter down only Uber’s financial document. If you don’t have a 2022 document but you do have one that discusses how Lyft compared to Uber or Uber’s financials in 2023, vector search will be able to pick that one up. While the details won’t be the most accurate for summarization, refinement using keywords and understanding of the text could help improve accuracy .

---


The topic of privacy and security is of utmost concern when dealing with AI, and specifically with generative AI. It has ethical, moral and legal implications, as can be seen by some of the recent data privacy lawsuits. One school of thought finds that open source is safer and more private. OpenAI, for example, is trained on open sources like GitHub and Wikipedia. Some of the privacy concerns arise when it comes to how these solutions are using the data users prompt. There were cases when users were able to receive other users’ private data, like phone numbers, as a response to their prompt. However, commercial solutions are not necessarily safer. When tuning your own model you own the data lifecycle and can control which data is going in. It’s important to ensure that data pipelines filter out private information and do not emit it. This call center demo example, uses an open source PII (personally identifiable information) recognizer to filter out private data patterns, like names, social security numbers, credit cards, etc. This ensures private data will not appear in the results when tuning or prompting the model.

---


To summarize information and sentiment from videos, it’s recommended to take a text classification approach. This is a similar approach to managing audio streams. First, generate as many features as possible from the video stream. To do so, ask questions about the stream. E. g, ‘What is the sentiment of the video? ’, ‘What object is shown? ’, or ‘How many people are in the image? ’. Then, generate a table of metadata based on that information and structure the information in a database. Finally, look up the information and format it into text. Note: We don't recommend running queries that examine all the pictures, since this is not a cost-effective process.

---


There are a number of use cases that require searching for and summarizing dynamic data. Examples include news websites, which require searching for news articles, or recruitment platforms, which require searching for candidates’ CVs. The basic way to manage this need is to take the document as is and index it. However, the resulting promo may not be the best one. Using associations, i. e vector search, which is based on seeing a familiar word, might also result in hallucinations. Instead, it is recommended to perform data preparation: processing, structuring and keyword searching, based on the document’s structure. Documents have a certain structure: a header, a body, tables, etc. Identify this structure. Then, determine which sections in the document highlight information you want to serve to your users and label them or add as a question. The paragraph within that section would be the answer. This way, the LLM knows this section is the relevant one. Understanding the document structure is very important, because it reduces the chances of hallucinations and saves resources. If the data is being regularly updated in a dynamic way, it’s important to also version the data, just like in data lakes and feature stores. Be wary of rushing to build a demo and abandoning engineering practices. When versioning, it’s important to refer to which version of the data you are using in the index search with labels and keyword searches, perform rolling upgrades, conduct metadata-based filtering, and more.

---


Collaboration is key for aligning on intents before implementation, to ensure the best results for the problem we are trying to solve. Collaboration does the glue work to ensure a consistent and streamlined workstream between data scientists, data engineers, designers, etc. For example, when implementing a human-centered approach for GenAI, joint decision making and sharing of practices and technologies helps ensure the development of the right prompts that answer the customers’ needs and ethical requirements. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


Training or even tuning a model requires lots of computation. You might even need distributed frameworks. There are two main distributed frameworks that are used in this space. Horovod, which interacts with the high-speed MPI layer for messaging underneath while running the decision training over TensorFlow or PyTorch. The second popular framework in space is Ray, which is also distributed Python. It allows you to distribute the workload and chart your model, whether it's based on the data or based on the model scaling across multiple computations. There are also frameworks that address serving at scale. An LLM like GPT-3 requires a significant number of GPUs, let’s say anywhere between four and eight. In terms of the budget, they require $20,000 - $30,000 per month for serving. Smaller frameworks like OpenLLaMA may also perform well if you teach them enough knowledge, requiring only one or two GPUs. There are also frameworks that know how to partition your model into multiple GPUs. Currently, there is a lot of research around how to build distribution more efficiently both for the serving and the training aspects. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


One method is to develop a second model to examine the first model and evaluate its response. But this space and technology are constantly evolving, so there is currently no technological market leader or established players that can evaluate responses. This is expected to change as the space matures. Until then, it is recommended to experiment as much as possible while involving humans to validate. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


A human-centered approach is the cornerstone of ethical AI. To implement this approach: Start by outlining the risks you want to avoid in terms of bias, transparency, explainability, fairness, toxicity, hallucinations and all the other dimensions that make up Responsible AI.   Define the metrics you will use to measure the presence of these risks. Measure the developed models against these metrics to ensure their reliability and trustworthiness. For example, is the generated content compliant with the bias metrics. You might develop custom-made algorithms that provide a layer of explainability about the models’ output. You can even develop an analytical engine that monitors the ML pipeline and the compliance with these metrics. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


There have been cases where chat information was leaked, which makes it all the more important to implement guardrails and involve the legal team. This will help minimize the risk of a privacy breach. As a general rule of thumb, it’s recommended to not share private data with LLMs since it might accidentally leak or get exfiltrated. This is especially important when running a POC with a public LLM. When the business is hosting an internal LLM, considerations might change. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


Fine-tuning is the process of taking a pre-trained model and training it further on a smaller, domain-specific dataset. This additional training helps the model adapt to the nuances and requirements of the specific task or domain. Fine-tuning helps transferring the general knowledge learned from the foundational training on to specific tasks or domains, to enhance the model's performance on those tasks. Embedding a foundational model means integrating or incorporating the pre-trained model into another system or application, but without significantly altering its learned parameters. This means using the model's capabilities as they are. For example, the model could be used for generating text, answering questions, or any other task it was originally trained for. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


Most of the MLOps steps are focused on structured data. To adapt them to LLMs, the steps need to be adapted. This includes, for example, embeddings, tokenization, data cleansing, and more. In addition, steps like evaluation and monitoring are more complicated and require more innovative thinking. For example, a drift analysis requires a histogram analysis and histogram comparison. This could require text analysis, conversion to numeric values, clustering or another type of logic to compare the training or the expected results with the actual results. Interested in learning more? Check out this 9 minute demo that covers MLOps best practices for generative AI applications. View this webinar with QuantumBlack, AI by McKinsey covers the challenges of deploying and managing LLMs in live user-facing business applications. Check out this demo and repo that demonstrates how to fine tune an LLM and build an application.

---


Organizations that do not have an external dataset for training their models or want to have a model that is trained on their own data, can use prompt engineering or fine tuning. Prompt engineering means feeding a model with engineered requests (prompts), which include specific content, details, clarifying instructions, and examples. These prompts guide the model into the expected and most accurate answer. For example, ChatGPT is currently being trained on prompt engineering. Prompt engineering significantly improves model performance and tailoring of AI outputs, ensuring they align with desired outcomes and ethical guidelines. It also simplifies user interactions by providing clearer instructions and racks up less computational costs over fine-tuning, since optimized prompts result in more accurate results with fewer resources. While it’s certainly possible to do prompt engineering on a 3rd party model, it’s critical to understand the challenges of doing it that way: Inference performance Prompt size is usually limited Versioning sometimes changes the responses  Company data is publicly exposed Another viable option is fine-tuning. Fine-tuning involves taking a pre-trained model, which has already learned from vast amounts of data, and further refining it on a specific task or domain. The model is exposed to labeled data related to the target task, allowing it to adapt and specialize its knowledge to better perform on that particular task. This process is highly beneficial as it enables the transfer of general knowledge acquired during pre-training to more specific applications. Fine-tuning also helps enhance the model's performance, boost accuracy, and ensure...

---


Training and deploying LLMs can be a costly activity. This is because training and deploying LLMs requires substantial computational power, including high-performance GPUs. These computing resources are used for running multiple iterations of complex algorithms on massive datasets when training, as well as for fine-tuning and for deployment. The cost of renting or owning these resources, along with the electricity required to run them, contributes to the overall expense. Training a foundation model is so expensive–GPT-4 was trained on a cluster with 25k GPUs for over a month, at an estimated $10M–so it’s really not feasible for any but a select few technology companies. However, there is a large and growing landscape of open and commercial LLMs that organizations can leverage and customize. Customizing an LLM–whether via prompt engineering or fine tuning and transfer learning–still comes with significant costs. MLRun and Nuclio are two open source solutions that can help optimize these customized LLM deployment costs. MLRun is an open source MLOps orchestration framework. MLRun can help optimize training costs by using state of the art algorithms and automated technologies, like quantization and distributed training. Quantization enables using smaller GPUs, which are more cost-effective, and distributed training allows faster training or even enables the training itself if the user doesn't have access to large enough GPUs. To optimize serving, MLrun leverages Nuclio. Nuclio automates the data science pipeline with serverless functions. By leveraging those serverless advantages, Nuclio auto-scales and employs on-demand resource utilization automatically at each step of the LLM...

---


MLRun can run as easily on Amazon SageMaker as it does on a local computer. In fact, it is environment-agnostic. For AWS users, the easiest way to install MLRun is to use a native AWS deployment. Here's how to do that. MLRun operates with a server side and a client side. The client side can run on everything, including AWS SageMaker, Azure ML, any Python, any Notebook, and more. MLRun can also run on Kubernetes Minikube, with containers, with Docker Compose, on EKS, on a cloud Kubernetes environment, and more. The only step required is to run pip install and configure environment variables. MLRun turns the requirements into a server function that uses auxiliary services. For running MLRun with Amazon SageMaker, MLRun provides two options: Using MLRun against workloads that will run on Amazon EKS  For building a workflow around SageMaker services like SageMaker Autopilot. For more on all the ways Iguazio works together with AWS, including solution briefs, demos and more, check out the partner page. https://www. youtube. com/watch? v=ensWP77Yayo

---


Seldon is an MLOps solution but it is not a serverless technology. Users will need to provide Seldon with code, containers, YAMLs, etc. Then, Seldon can be used as a framework for launching them. It will build the entire service, including auto-scaling and APIs. MLRun, on the other hand, orchestrates ML and MLOps end-to-end, including pipelines and serving capabilities. In addition, MLRun has metadata services for saving objects, like models, datasets, features, etc. When training, MLRun automatically builds a survey with the file and information about the data schema, statistics, for drift analysis, parameters, and more. In other words, MLRun supports glueless integration with the serving layer and the monitoring layers without having to write a single line of code.

---


Yes. Here’s how it works: In MLRun, projects are the fundamental unit for working with the platform. First, users add or edit code and configurations in their project. Then, they run it and debug locally. Once ready, the code and configurations are pushed into a source repository (git) and tagged with labels. Finally, the code and configurations are loaded so they can run on development or production clusters. Users can load different versions of their projects, like development, staging and production. MLRun also integrates with CI systems. These CI systems enable automatically pushing code to the next required step. For example, to a development environment, to the testing phase, into a deployment flow with canaries, etc. Users can configure the CI to promote models to any environment or account. For more information from the docs on getting started with AWS and MLRun, see here. For more on all the ways Iguazio works together with AWS, including solution briefs, demos and more, check out the partner page.

---


MLRun is made up of multiple services: a UI, orchestration, security, RBAC, SSO, and more. These services are provided through multiple types of components, with operators being one of them. MLRun uses operators when launching services like a Spark job or a real-time serving function. These operators take the requests and turn them into live services. MLRun supports eight different operators: Nuclio for real-time, operators for multi-stage graphs, an operator for online multi-state graphical serving, a Spark operator, a Dask operator, and more. These operators are pluggable, which provides flexibility for adding more in the future. Operators are deployed as Helm charts, but can also support cloud formation.

---


MLRun supports GPUs. When launching jobs or real-time functions like serving and data processing, you can turn on the GPUs. You can also specify how many CPU, GPU or CPU ratios are required. When building underlying serverless components, MLRun knows how to automatically configure GPUs for optimal usage. For example, when running a training job and requesting a GPU, MLRun will configure CUDA drivers automatically, enabling seamless transition from using a GPU to not using a GPU with a single flag. In this mask detection demo, you can see how to seamlessly move from distributed work on multiple GPUs to running locally without the GPU with the flip of a parameter.

---


A batch is a group of instances from the dataset. Batches are fed into the different phases of the MLOps pipeline, like the training phase, for processing. Batches are used for use cases that do not require real-time or online data, which makes it more resource efficient and simple to use. You can practice how to run batch inference with MLRun with this tutorial.

---


Real-time streaming is the process of collecting and ingesting data and processing it in real-time to answer business use cases. In MLOps architectures, streams store sequenced information, or records of data, that can be used for production and consumption. Consuming the data from the streams in real-time, enables running applications that require online data, like e-commerce or fraud detection. You can check out how to run a real-time pipeline with MLRun with the following demos: Model monitoring and drift detection (deploying a model to a live endpoint) Advanced model serving graph (deploying the graph as a real-time serverless functions) Distributed (multi-function) pipeline example

---


MLRun is an open source MLOps orchestration tool at the core of the Iguazio MLOps Platform. MLRun integrates data preparation, model development, model and application delivery, and end-to-end model monitoring. MLRun provides an open and pluggable architecture, so you can build a pipeline in MLRun that launches the jobs in Azure ML or SageMaker (and soon, in Databricks). Here are some examples of ways you can use other services with MLRun: Develop on SageMaker, Deploy with MLRun Develop on Azure ML, Deploy with MLRun Once the model is generated, MLRun can build it into an application pipeline. With MLRun’s pluggable architecture, you’re not locked into a single environment.

---


There is some overlap between MLFlow and MLRun, but they have totally different goals. MLRun is an end to end orchestration layer for ML and MLOps. It’s not primarily a tracking system, though it does offer that functionality. MLFlow is a way to track your experiments, a component in the experimentation phase. There are also some ways to define metadata in MLFlow. However, there are many parts of the MLOps lifecycle that MLFlow doesn’t cover, like applying automation, automating serverless functions, running jobs, model monitoring, preparing data logic, and so on. Airflow is a great tool for when you want to track experiments and show great charts showing experiment results but it’s not a tool that will save time getting to production. MLRun is for what we call AutoMLOps, where the entire operationalization process is automated. MLRun uses serverless function technology: write the code once, using your preferred development environment and simple “local” semantics, and then run it as-is on different platforms and at scale. MLRun automates the build process, execution, data movement, scaling, versioning, parameterization, output tracking, CI/CD integration, deployment to production, monitoring, and more. MLRun provides an open pluggable architecture, so you have the option to use MLFlow (or any other tool) for the development side, and then use MLRun to automate the production distributed training environment without adding glue logic.

---


The short answer is yes. Everything in MLRun (Iguazio's open source MLOps orchestration framework that sits at the core of the Iguazio MLOps Platform) is versioned: data objects, feature sets, feature vectors, and models. There are actually two different versioning schemes, like a git repository, with the shots or unique identifier of a version, as well as tags and labels. Objects can have a version, like production or development, and they can be changed. There is also the aspect of labeling for the training set and there are many different tools (that integrate with MLRun) that can do the labeling manually. Streaming data requires an automated approach, where the gets labeled as it is ingested. For more on different approaches to automated labeling, check out chapter 2 in the book Implementing MLOps in the Enterprise by Yaron Haviv, Iguazio co-founder and CTO and Noah Gift, MLOps expert and author, from O’Reilly

---


Overlapping areas of responsibility (as in cases where multiple roles are expected to do things like data preparation or model building) is a common problem on ML teams. By defining a “contract” between the different components of MLOps, as well as creating automations, you eliminate a lot of those challenges. Another way to make teams more efficient is for everyone to work on the same platform, using the same APIs and metadata databases. For instance: If a data scientist has built some data preparation code in a Jupyter notebook, it can be automatically turned into a pipeline that uses Spark. Now all you need is data from a data warehouse and it will use the same business logic, just on a scalable machine. You eliminated a lot of that friction between roles using automation. When a data scientist builds a feature, he/she doesn’t need to care about validation policies and biases. Instead, his contract is to define those features, and then move on to the training. The data engineer can take that definition and continue working on it—validating or cleaning the data in a different manner. One member of the team can build a serving pipeline just to test it out to see that it works, and then another member in charge of the production side can update it to work with GPUs, or spot instances.

---


Nuclio and MLRun are two open-source technologies that the Iguazio team maintains. Nuclio is a serverless platform, and MLRun is an MLOps orchestration framework. These two open-source technologies are major components that run under the hood and power our enterprise offering, the Iguazio MLOps Platform. Read all about MLRun in the documentation here and join the Slack for community support here. Read about the technology behind the Iguazio MLOps Platform here, and the Iguazio documentation here.

---


A data warehouse is for structured reporting data that is typically used for reporting data with consistent business data points (for example, metrics like quarterly sales). The structure of data in a data warehouse is well known, and it holds critical business data for reporting purposes. In a data warehouse, lots of work goes into structuring and cleaning the data upfront, so that it’s selective and useful. A data warehouse housed in a costly proprietary system, like Oracle or Terradata, so it makes sense to store the most critical business data here. A data lake offers massive storage for a much cheaper price point, so you don’t need to know or care about the structure or content of the data. The approach with a data lake is to “dump it in, and deal with it later”. This kind of setup is for any kind of data science process, where the data could potentially hold some value, but exploratory analysis is required to uncover it. A data lake requires very little work up front, and heavy data engineering—processing, running transformations and calculations, etc--to extract the value later. It's worth noting that data infrastructure changes along with the maturity of data use cases in organizations. Data warehouses are primarily for business intelligence, and data lakes are built once some kind of data science work has begun. Once multiple models need to be built and maintained, the next step of maturity is a feature store with advanced data transformations, where data scientists and...

---


Batch and stream processing are two types of methods used to process data, which is one of the steps in feature engineering. The processed data can be used in features for generating either real-time or batch predictions. In Batch processing, historical static data is processed in batches for use in features. Batches might run at scheduled intervals, or run when compute resources are available. With batch processing, heavy feature computations can be performed offline, so that it’s ready for fast inference. However, features do become stale, as changes in the real environment shift over time. It’s important to set up a drift-aware monitoring system for these features that includes a retraining step. In Stream Processing, predictions are made on real-time inputs with near real-time or streaming features for a given entity. Predictions made with stream processing and therefore streaming features can improve the quality of predictions by adding more valuable data to the model. For instance, a recommender system with streaming features can make use of a user’s recent website behavior, combined with data like real-time inventory or purchase history. Stream processing is quite a bit more complicated to embed into an ML workflow, and requires a real-time feature store, mature streaming infrastructure, and an efficient development environment so data scientists and ML engineers can work together to validate new streaming features.

---


In Static Deployment, the model is trained offline, with batch data. The model is trained once with features generated from historical batch data, and that trained model is deployed. In static deployment, model training is done on a local machine and then saved and transferred to a server to make live inferences. All predictions are precomputed in batch, and generated at a certain interval (for example, once a day, or once every 2 hours). Examples of use cases where this type of deployment is appropriate are situations like content-based recommendations (DoorDash’s restaurant recommendations, Netflix’s recommendations circa 2021) In Dynamic Deployment, the model is trained online, or with a combination of online and offline features. Data continually enters the system and is incorporated into the model through continuous updates. In this deployment scenario, predictions are made on-demand, using a server. The model is deployed using a web framework like FastAPI or Flask and is offered as an API endpoint that responds to user requests, whenever they occur. Examples of use cases where this type of deployment is used are real-time fraud detection and prevention, delivery or logistics time estimates, and real-time recommendations.  

---


Distributed computing tools are a great way to run training on very large datasets. Many ML use cases require big data to train the model, which complicates development and production processes: In cases where data is too large to fit into the local machine’s memory, it is possible to use native Python file streaming and other tools to iterate through the dataset without loading it into memory, though this method can be slow going, since jobs are run on a single thread. Parallelization can speed up the process, but you’re still limited by the resources in your local machine Distributed computing tackles this challenge, by distributing tasks to multiple independent worker machines, each handling a part of the dataset in its own memory and dedicated processor. By using distributed training, data scientists can scale their code on very large datasets to run in parallel on several workers. This solution surpasses parallelization where jobs run in multiple processes or in threads on a single machine. There are two main options for transitioning to a distributed compute environment: Spark: A very mature and popular tool for data scientists who prefer a more SQL-oriented approach, especially in enterprises with legacy tech stacks and JVM infrastructureDask: The go-to tool for large-scale data processing for those who prefer Python or native code. It has seamless integration with common Python data tools that data scientists already use. Dask is great for complex use cases or applications that don’t neatly fit into the Spark computing model. More...

---


Moving your pipelines from batch to real-time is a complex endeavor. In many cases, the pipeline needs to be redesigned with a new tech stack that supports event-based architectures. Before you even contemplate moving your pipeline online, you must determine whether the data that supports the pipeline can also be moved to (near) real-time (streaming tech is the most popular right now). If your batch pipelines are supported by daily ETL that takes hours to process, even if you move the pipeline processing online, the data doesn’t support that. Start with the data: Where does it originate? Can you get a hold of the data when it originates, such as when a user clicks a button on a mobile app? Solving these issues can involve changing source applications and using streams. If it can be done, the battle is halfway won! Now that you have access to data streaming from the source, you need a way to run that data through your pipeline logic at scale. Iguazio's serverless open-source framework Nuclio allows you to create complex pipelines with many different event-based triggers. It also scales horizontally to fit the demand of the workload.

---


When automating a pipeline of any kind, whether it be ML pipelines like model training or data-focused pipelines such as feature engineering, the basic first task is to remove and/or replace any step in the pipeline that requires a human to do the work. There are instances where human logic is ok in a pipeline, such as “Review” and “Approve”, but generally we want everything else in the pipeline to be able to execute without help from us. There are many tools out there for creating pipelines and job workflows. Here at Iguazio, we use Kubeflow Pipelines. Next, you need to design the pipeline in such a way that it’s robust and resilient. Think about how to answer questions like these: If my pipeline fails, can I easily restart it without manual fixes? If my pipeline executes multiple times in a row, do I still get the desired results? Is my pipeline dynamic in how it handles new data, parameters, scale, etc? If my pipeline routinely has issues/risks, what are they and how do I mitigate them? Before involving any CI/CD framework, you should be able to manually “trigger” the pipeline and answer these questions through some basic tests. Once you are satisfied with the pipeline, it’s time to automate it. What are the conditions under which this pipeline should run? If the answer is something along the lines of “every night at X time”, then you can accomplish that with a simple scheduling mechanism, which is often times part...

---


The performance of a machine learning model degrades over time, absent intervention. This is why model monitoring is an important component of a production ML system. When a machine learning model’s predictions start to underperform, there can be several culprits. After ruling out any data quality issues, the two usual suspects are data drift and concept drift. It’s important to understand the difference between them, because they require different approaches. In data drift, the input has changed. The trained model is no longer relevant on the new data. In concept drift the data distribution hasn't changed. Rather, the interaction between inputs and outputs is different than before. ֵIt practically means that what we are trying to predict has changed. A classic example is spam detection: over time, spammers try new tactics, so the spam filters need to be retrained to react to these new patterns. For a deep dive on how to build a drift-aware production ML system, check out this blog. For a super-simple explanation of these concepts, check out Meor Amer's brilliant illustration of this idea: Source

---


Self-supervised learning is an evolving technique of helping ML models to learn from more data, without the bottleneck of human-labelled datasets. In this technique, the model predicts any hidden part of the input from an unhidden part of the input. This is similar to how you’re able to predict the missing letters in Wheel of Fortune: your brain fills in the hidden letters based on the clues in the letters you’re shown. By contrast, supervised learning is a technique where the ML model learns from labelled data sets, typically for tasks like classification. One benefit of supervised learning is that its performance can be measured during training, to assess how well the model has learned. So why would you choose one over the other? Machine learning models need to learn from massive amounts of good-quality labelled data. Finding or creating labelled data is a huge bottleneck, because it takes so long and costs so much—for instance, to teach a model to recognize a picture of a dog, a human has to label images as “dog” or “not a dog”. Practically speaking, it’s impossible for humans to label everything in the world (especially unstructured data), and there are some tasks for which there is simply not enough data, meaning any potential AI system would be limited by a small training set. Self-supervised learning addresses these limitations, allowing ML teams to scale the research and development of ML models at a low cost.

---


You are basically asking for model serving or a way to manage and deliver your models in a secure and governed way to production. There are a few things you need to think about: How will my models be managed? How will my models be delivered (served) for inferencing? Do I need real-time or batch level delivery? In its simplest form, you store or deploy the trained model to a remote repository known as a model server. Then at runtime, you retrieve the model, pass features (inputs) into it and predict. There's a lot of value in this simple model. Firstly, your models are stored in a central repository which provides governance, share-ability, versioning and reusability. It should be as easy as a few function calls. Secondly, retrieving the model should also be as easy as a single function call. However, you must ensure the appropriate protocols are supported and are secure. A great way to accomplish this is using MLRun Serving Pipelines. Using MLRun Serving uses the Nuclio real-time serverless framework for the pipelines.

---


In a nutshell, model monitoring allows a data scientist or DevOps engineer to keep track of a machine learning model after it has gone into production. This includes keeping track of resources, latency, invocations, and data/concept drift.  Using this set of tools, we can define several types of use cases:  Keep track of how often and how quickly your model is being used.  If a certain time of day is much busier than others, this will become immediately apparent and allow your team to plan as such. This is especially useful when used in conjunction with viewing the consumed cluster resources. This allows for overall capacity planning and may showcase some opportunities to configure auto-scaling behavior.   Calculate data and concept drift.  This will give insights on the incoming data as well as the outgoing predictions from your model – both of which can drift and can be indicative of problems.   Drift-based event triggering.  Based on a drift event, it is possible to start a re-training pipeline that will train and deploy an updated version of the model. This is known as continuous training. Additionally, it is also possible to start a statistical analysis of your training data vs the received live data to see where the differences lie.   These abilities are non-exhaustive but rather some of the many capabilities provided by proper model monitoring.  

---


After spending a long time developing and training our model, it’s finally time to go to production. But how do you still know if your model is still making accurate predictions in a week or month from now?  How do we know how many resources the model is using?   In short, model monitoring keeps track of your model after it goes into production. There are several facets of monitoring including:  Kubernetes resources: How many resources are we using (CPU, MEM, GPU)?  Latency: How long does inferencing take?  Number/Average Invocations: How often is our model being used?  Data drift: How different (statistically) is the incoming live data from the data we used to train the model?  Concept drift: Has the meaning (statistical properties) of our prediction target changed?   While all of these facets are important, some are easier to compute than others. Natively, Kubernetes reports resource utilization. It’s possible to get latency and invocation information from most model serving frameworks. However, data and concept drift are incredibly difficult to compute in real-time and on an on-going basis.   This is one of the cases where an MLOps platform such as Iguazio will aid in deploying your models with built-in model monitoring and dashboards.  

---


Where the feature store fits into the overall ML lifecycle depends on the functionality of the feature store.  Some feature stores are purely for storing and retrieving batch datasets.  With these kinds of feature stores, they would likely be used after an ETL job from a data engineer to store the newly transformed data or after a feature engineering job from a data scientist to store the newly created features.  From there, the end user can browse the feature store and see what is available to retrieve.  Because this feature store is purely for storing and retrieving features, it has more limited uses within the overall ML lifecycle.   Some more advanced feature stores support batch and real-time transformation engines, real-time feature serving, and model monitoring integration. This allows the feature store to be used throughout the entire ML lifecycle from data ingestion/transformation, mode training, model deployment, and model monitoring. Keeping all of these processes centralized in a single place allows for standardization across an organization as well as glue-less integration between different aspects of your ML lifecycle.  

---


Aside from the technical benefits of a feature store, one of the main benefits is organizational. In a typical enterprise ML team, the data engineers and data scientists have their own siloed workflows. As an over-simplification, data scientists will ask data engineers for new features and the engineers will provide them to the scientists. In this workflow, data scientists are decoupled from the data sources and are often unaware of problems providing and transforming data.  Additionally, data engineers are decoupled from the usage of the data and are often unaware of problems implementing and utilizing data.   In this context, a feature store will help both parties. It will allow both teams to gain a better understanding of the existing features, data sources, transformations, etc. It will also empower both parties to aid the other. For example, data scientists will have the ability to ingest from various data sources and perform transformations in batch and real-time. Additionally, data engineers will have the ability to see what features are currently available and retrieve them easily in batch and real-time.   In this way, organizations can utilize a feature store to ease friction of hand-off points between teams.  

---


The concept of an ML pipeline is an automated pipeline that can be created from steps that take a model to production. In each step, different pieces of code can run that handle different types of processing within that pipeline. In a typical pipeline, there is a need for different functions that handle actions like collecting data sets, cleaning and preparing the data, calculating the data, creating feature sets, running training processes, selecting the best model from the training process, deploying the model on an inference layer, and finally, monitoring the model. An operational pipeline needs to support running those processes in a scalable way, so it needs to run on a framework like Kubernetes, which enables you to scale up or down based on your load. This pipeline also needs to support running different frameworks, for example, you may want to run your data preparation using Spark or Dask, or when it comes to processing data in real-time, you may need to incorporate frameworks that can read streaming data, such as Nuclio or Spark streaming. Another thing that’s important when creating a pipeline is the ability to track and capture relevant matrix and logs so the user can easily compare between different runs and identify issues within the pipeline and the root cause of those issues. Businesses seeking to generate profit with machine learning will benefit from MLOps. Businesses that are already further along on their data science journey will be more likely to see the advantages of MLOps...

---


Robust Data Transformation and Real Time Feature Engineering A feature store provides a means for creating a feature list in a logical group. This ensures they can be ready for training or inference. However, a feature store is much more than just a catalog. It is also a data transformation service for creating features. Data scientists can easily create features using simple APIs. These APIs allow them to create complex functions, including aggregation, sliding window, joins custom functions, etc. By using these abstract APIs, data scientists can create complex features while reducing the dependencies on data engineers. Advanced feature stores can handle features for batch processing as well as real-time feature engineering. As a result, the same abstract API can be used for calculating features in real-time based on event streaming (e. g. streaming coming from Kafka, Amazon Kinesis, etc. ). Write Features Once Data scientists often create features while they prepare their models for training. However, once they need to take it to the operational production pipeline, they pass it over to the data engineers. Then, the data engineers need to write code in Spark or Java to make it ready for production. This process becomes much easier with feature stores. Through the feature store APIs, features can be reused for both training and online inference, without the need to rewrite the code again for production. Monitoring and Drift Detection Alongside the features definition and the code for creating them, feature statistics are also captured. Advanced feature stores are...

---


What is Data Preprocessing? Data preprocessing is the process of cleaning and preparing the raw data to enable feature engineering. After getting large volumes of data from sources like databases, object stores, data lakes, engineers prepare them so data scientists can create features. This includes basic cleaning, crunching and joining different sets of raw data. In an operational environment this preprocessing would run as an ETL job for batch processing, or it could be part of a streaming processing for live data. Once the data is ready for the data scientist - then comes the feature engineering part. What is Feature Engineering? Feature engineering is the creation of features from raw data. Feature engineering includes: Determining required features for ML modeAnalysis for understanding statistics, distribution, implementing one hot encoding and imputation, and more. Tools like Python and Python libraries are used. Preparing features for ML model consumptionBuilding the modelsTesting if the features achieve what is neededRepeating the preparation and testing process, by running experiments with different features, adding, removing and changing features. During the process, the data scientist might find out data is missing from the sources. The data scientist will request preprocessing again from the data engineer. Deployment to the ML pipeline How Do Data Preprocessing and Feature Engineering Relate? In preprocessing data engineers get and clean data from the sources to be used for feature engineering. Feature engineering is the part of creating the actual features.

---


The key components of a successful MLOps strategy revolve around having standards and best practices that will help you develop your ML service in a way that is almost production-ready from day one. It’s very important to have a production mindset even if you’re still in the development phase. This includes: 1. Tools & Frameworks Choose the right tools for your data scientists, data engineers, and DevOps. These tools and frameworks should be used in both development and production environments, in a way that will smoothen out the transition between development and production. 2. Building Blocks Build the right building blocks for: Data acquisitionFeature engineeringModel trainingModel servingModel monitoringGovernance For example, feature stores can help with feature engineering, which, in many cases, is considered the hardest task when building ML pipelines. Feature stores are a catalog of features that data scientists and engineers can leverage. With feature stores, data scientists can reuse features instead of creating duplicate features. They also empower data scientists to run complex feature engineering tasks using a simple and abstract API. 3. Strategy A strategy that covers all these steps. So when you start you can feel confident you have the right best practices and standards to complete each step in a way that reduces the time it takes to get those components to production. The state of mind should be that the development process should be production-ready from the get-go.

---


In the race to solve business problems, more companies have invested considerable capital into becoming data-driven. The hiring cycle is beginning to prioritize data scientists, data engineers, and DevOps engineers. What glues this team and the ML solutions pipeline together? MLOps platforms. Pros: Remove silos with collaborative spaces and feature sharing, scale with continuous deployment, monitor for drift at scale Once the data team is complete and some models are trained, companies begin to sense a problem. Silos between data scientists, engineers, and DevOps friction the path to deployment. The path to production is long and requires patience and stamina. As a result, valuable models get stuck in production. MLOps platforms like Iguazio allow for collaboration through their UI and feature store, continuous integration and deployment, and model monitoring. Cons: Initial costs The only aspect that MLOps could be considered a “con” is the costs outside of long-term cost-benefit analysis. That is, introducing MLOps to your firm might be expensive if you think short-term. However, if you project the benefits of driving your decisions with data, then the only thing you should consider is whether you build or buy your MLOps solution.

---


On a high level, the concept of pipelines in ML refers to a way of linking sequential components of the workflow of an ML project, and their relationships. In ML projects, it is challenging to keep track of when and where steps like data prep, training, and monitoring take place, so pipelines serve as a step-by-step map that describes what needs to happen and when. Working with pipelines creates easily composable, shareable, reproducible, and stitch-able ML projects. Kubeflow pipelines in particular is a set of services and UI enabling the user to create and manage ML pipelines. Users can write their own code or build from a large set of pre-defined components and algorithms contributed by teams at companies like Google, IBM, Amazon, Microsoft, NVIDIA, Iguazio, and others. Much like a function (ingesting inputs, parameters, and producing outputs), each Kubeflow component is Python code that is packaged in Docker images that executes one step within the ML pipeline. Kubeflow components will launch one or more Kubernetes pods for each step on your pipeline. Kubeflow can also run on Nuclio, a high-performance serverless platform that runs over Docker or Kubernetes and automates the development, operation, and scaling of code. The benefits of using Kubeflow are: Scalability: Easily spin up more resources when needed, and release them when you don’t. Composability: Each step is independent, which simplifies the orchestration of the whole pipeline. IE: you can use many different ML-specific frameworksPortability: Easily compose each step of the pipeline in one place without...

---


In short, any business with a data science team that produces ML models that address business operations. The phrase “data-driven” is by now a cliche in the business world. Business decision-makers keep an eye on the bottom line with the help of internal data infrastructures. Depending on the industry, the method by which this value is extracted can look very different. In general, any business that collects and stores data can also train a model. Depending on the use case, models can deliver predictions and insights, and detect risks with data generated by the business.   MLOps becomes relevant when the model creation process is iterated, automated, and monitored so that the model continues to produce relevant results and business value as data continues to flow.

---


To maintain the accuracy of an ML model in production, and detect drops in performance, it can sometimes be useful to create custom metrics that are quite specific to the product.   Coming up with the right mix of metrics to monitor can be challenging. To get started, consider these two questions: 1. What are the expectations of the end user? Think of the metric as a basic user story: For a healthcare mobility use-case, a user story might be something like: “As a hospital worker who needs to triage patient care, I would like the most time-critical patient cases to be easily accessible, and therefore placed high on my screen. ” What metrics have an impact on your end-user’s perspective? 2. What is a bad user experience for your use case?   Instead of looking at ideal or typical experiences, consider some edge cases. What happens when your service delivers a bad user experience? This can be instances where the model delivers a fallback response, or a low certainty response, or even an empty response. Your model monitoring should be able to catch these instances. Bugs in any number of code locations can and do happen regularly at every enterprise. For ML models serving critical functions, real-time metrics will ensure that any drop in performance can be addressed ASAP.

---


The concept of an ML pipeline is an automated pipeline that can be created from automated steps that take a model to production. In each step, different pieces of code can run that handle different types of processing within that pipeline. In a typical pipeline, there is a need for different functions that handle actions like collecting data sets, cleaning and preparing the data, calculating the data, creating feature sets, running training processes, selecting the best model from the training process, deploying the model on an inference layer, and finally, monitoring the model. An operational pipeline needs to support running those processes in a scalable way, so it needs to run on a framework like Kubernetes, which enables you to scale up or down based on your load. This pipeline also needs to support running different frameworks, for example, you may want to run your data preparation using Spark or Dask, or when it comes to processing data in real time, you may need to incorporate frameworks that can read streaming data, such as Nuclio or Sparks streaming. Another thing that’s important when creating a pipeline is the ability to track and capture relevant matrix and logs so the user can easily compare between different runs and identify issues within the pipeline and the root cause of those issues. Businesses seeking to generate profit with machine learning will benefit from MLOps. Businesses that are already further along on their data science journey will be more likely to see the advantages...

---


MLOps (Machine Learning Operations) is a combination of Machine Learning and DevOps principles. It uses DevOps concepts to manage the entire lifecycle of developing and deploying machine learning models.  When developing a new ML-based service, the data science element is just the first step. Making the model operational in the live environment is where the complexity lies, and that’s what MLOps addresses. With MLOps, the process of deploying a model is accelerated and streamlined, minimizing the level of DevOps effort, reducing time to market and improving model accuracy. MLOps seeks to increase automation and achieve a CI/CD approach to releasing production models.   MLOps builds and automates the entire machine learning lifecycle. These steps are:  Data collection and processing Feature engineering Model training Model Deployment Model Testing Monitoring / drift detection Drift remediation  Governance is another important element of MLOps. Enterprises need tools and processes to ensure data quality and security, enable explainability, and allow appropriate data access for auditing purposes.  

---


Sub title Lorem ipsum dolor sit amet, consectetur adipiscing elit. Etiam magna est, efficitur ut sem in, condimentum finibus enim. Nunc ac ex lacus. Phasellus semper mauris vitae erat bibendum, ut accumsan leo auctor. Suspendisse gravida at lorem sit amet rutrum. Vivamus nec turpis finibus, pharetra nunc vel, vulputate tortor. Vivamus lectus dolor, rutrum imperdiet accumsan nec, aliquet non tortor. Donec tempor metus dictum finibus venenatis. Etiam ac semper tellus. Donec ante urna, placerat et condimentum sit amet, imperdiet ut odio. Vestibulum scelerisque nunc enim, a auctor orci finibus eu. Etiam in velit leo. Mauris venenatis id tortor sit amet semper. Phasellus vestibulum dui non enim maximus posuere. Nulla facilisi. Ut at nisi egestas, aliquam erat sed, porttitor augue. Cras convallis ante sed ultricies porta. Mauris imperdiet ipsum vel ex lacinia hendrerit. Vivamus vitae lorem elementum, mattis ante sed, scelerisque augue. Donec id posuere dolor, vel hendrerit massa. Aenean placerat enim quam, nec sagittis est mollis facilisis. Sed turpis magna, porttitor at accumsan sed, vehicula eu velit. Morbi viverra augue sit amet enim aliquet, at molestie est efficitur. Cras vel varius sapien. Suspendisse volutpat tellus et bibendum ullamcorper. Proin orci quam, porta sed viverra ut, dictum eu felis. Vivamus vestibulum diam in velit commodo, ac hendrerit turpis fermentum. Mauris vel ex eu diam sollicitudin tempus sit amet non nunc. Morbi vulputate ligula faucibus, convallis sapien at, efficitur mi. Morbi viverra dui sit amet ultricies faucibus. Pellentesque congue auctor tempus. Morbi nec dictum dolor. Integer maximus tellus sit amet ante convallis,...

---


---