Redis Introduces Redis Feature Form, an Enterprise Feature Store for Production Machine Learning

Redis introduced Redis Feature Form, a managed feature store platform meant to help machine learning teams bring features into production across training and inference workflows. The update is aimed at improving control and consistency when teams define, orchestrate, version, and serve features across environments.

Redis said production ML teams often face operational challenges when models move from development to deployment, including drift, brittle pipelines, and governance gaps. Feature Form is described as a governed system intended to replace fragmented, homegrown workflows so training and serving stay in sync while managing feature pipelines in production.

Redis Feature Form is built for governed feature management across training and inference workflows. Redis said the latest release adds workspaces for multi-tenancy, fine-grained job control, atomic DAG updates, enhanced RBAC and security, simplified deployment, and a redesigned dashboard. Redis also said the platform manages graph-level changes atomically and includes audit logs, mTLS, encrypted internal transport, and improved secret-provider capabilities.

Redis said it acquired Featureform in October 2025, and Feature Form is the newest integrated version of that technology within Redis’ real-time data platform. The release also described unified batch and streaming pipelines, including support for tiling, backfills, and incremental updates, along with a leaner two-service deployment model. “Feature Form expands Redis’ role in the ML stack from a fast serving layer to a more strategic platform for production ML,” said Rowan Trollope, CEO of Redis.