Feast
feast.devThe Open Source Feature Store for Machine Learning
AI Toolsfeature-storemachine-learningmlopsopen-sourcepythondata-engineeringreal-time-serving
About
Feast is an open-source feature store that enables ML teams to define, manage, and serve features for both model training and real-time inference. It provides an offline store for historical feature extraction and an online store for low-latency serving in production systems. Feast integrates with existing data infrastructure and offers a Python SDK, CLI, and UI for feature management.
Problem
ML teams struggle to consistently manage and serve features across training and real-time inference while avoiding data leakage and infrastructure complexity.
For
ML platform engineers, MLOps engineers, data scientists, and data engineers building production AI/ML systems
How it works
Feast connects to existing offline and online data stores, allows teams to define features via a Python SDK, and serves them consistently for model training and low-latency production inference.
Business model
open-source
Status
launched