Feast

Feast (Feature Store) is an open-source feature store focused on solving training/serving skew: the mismatch between features computed for model training and features computed for real-time inference. Teams define features as code against existing data sources (warehouses, data lakes, streaming systems), and Feast materializes them into an online store (Redis, DynamoDB, and others) for low-latency serving while keeping an offline store for point-in-time-correct training datasets. It does not compute features itself the way a transformation tool does; it registers definitions and manages retrieval. Feast is governed as a Linux Foundation AI & Data project rather than owned by a single vendor, and is typically self-hosted on a team's own infrastructure or Kubernetes cluster. It suits teams that already have a feature-engineering pipeline and need a lightweight, vendor-neutral registry and serving layer rather than a full managed ML platform.