MLOps Live

Join our webinar on LLM Validation & Evaluation - Tuesday 26th of March 2024 - 9am PST / 12 noon EST / 6pm CET

Integrated Feature Store

Easily engineer online and offline features, share them across teams and ML applications with minimal development and integration effort

Works with

A Central Hub to Produce,
Share and Monitor Features

Features are properties that are used as inputs to a machine learning model. For instance, a recommendation application might use the total amount per purchase or product category as one of its many features. Generating a new feature, called feature engineering, takes a tremendous amount of work. The same features must be used both for training, based on historical data, and for the model prediction based on the online or real-time data. This creates a significant additional engineering effort, and leads to model inaccuracy when the online and offline features do not match. Furthermore, monitoring solutions must be built to track features and results and send alerts of data or model drift.

The Iguazio integrated feature store, at the heart of its data science and MLOps platform, solves those challenges. Accelerate the development and deployment of AI applications with automated feature engineering, improved accuracy, feature sharing and glueless integration with training, serving and monitoring frameworks.

 

The Only Feature Store with
Advanced Data Transformation

Transforming Data into Advanced Offline and Online Features

The Iguazio feature store is the first commercially available production-ready feature store which is part of an integrated and glueless data science and engineering solution. The Iguazio feature store automates and simplifies the way features are engineered, with a single implementation for both real-time and batch. High-level transformation logic is automatically converted to real-time serverless processing engines which can read from any online or offline source, handle any type of structures or unstructured data, run complex computation graphs and native user code. Iguazio’s solution uses a unique multi-model database, serving the computed features consistently through many different APIs and formats (like files, SQL queries, pandas, real-time REST APIs, time-series, streaming), resulting in better accuracy and simpler integration.

Build Features Once and
Plug Them Anywhere, Seamlessly

Building a Feature Vector From Features

The Iguazio feature store is a centralized and versioned catalog where everyone can engineer and store features along with their metadata and statistics, share them and reuse them, and analyze their impact on existing models. Iguazio’s integrated feature store plugs seamlessly into the data ingestion, model training, model serving, and model monitoring components, eliminating significant development and operations overhead, and delivering exceptional performance. Users can simply collect a bunch of independent features into vectors, and use those from their jobs or real-time services. Iguazio’s high performance engines take care of automatically joining and accurately computing the features.

 

Model Monitoring, Governance,
and Automation, Built-in

Real-Time Features and Drift Detection

The unified online and offline feature store provides next-level automation of model monitoring and drift detection, training at scale, and running continuous integration and continuous delivery (CI/CD) of ML. Features are stored along with their data quality policies and auto generated online and offline statistics, to automatically detect model drift, inaccuracy, and alert the users or initiate automated re-training workflows.

 

AutoML

Robust Data Transformation

Create complex feature engineering processes with a built-in robust data transformation service, including feature aggregations with sliding windows, dozens of pre-built transformations, or your custom logic in native Python code. With a simple API and SDK, data scientists can easily create features without requiring long data engineering cycles.

Churn Reduction

Feature Catalog

Share, search and collaborate on features, evaluate features with detailed statistics and analysis, and see how features correlate to both data sources and models with an easy-to-use user interface.

Mitigate Risk

Integrated Data and Model Monitoring

Capture the feature statistics in real time, enabling drift detection based on actual data drift. The Iguazio feature store is fully integrated with the rest of the MLOps Platform, with features like concept drift monitoring and feature monitoring out of the box.

Real-Time Graph

Real-Time Feature Engineering

Features are developed once for offline and real-time; no extra work is needed. The feature transformation pipeline calculates features in real time based on incoming events or streams, and serves the results at millisecond level latency or pushes them directly into a stream.

Data Governance

Keep the data lineage of a feature, with the tracking information capturing how the feature was generated, critical for regulatory compliance.

Deployed Anywhere

Hybrid Deployment

Build models using the integrated feature store and deploy them in hybrid multi-cloud environments, on-prem or at the intelligent edge: anywhere your application lives.

Benefits

Faster Deployment of AI

Faster Deployment of AI

Improved Model Accuracy

Improved Model Accuracy

Advanced Data Governance

Advanced Data Governance

Collaborate and Re-Use

Collaborate and Re-Use

Learn More

Data Science Platform Tutorials

Platform Overview

Get started with a video introduction to the Platform

Data Science Platform Documentation

Documentation

Access overviews, tutorials, references and guides

Build, Share and Manage Features Across the Organization

Eliminate silos, automate complex online and offline feature engineering and share features across teams and projects with the Iguazio feature store