Hopsworks
Unified feature store and AI platform from raw data to production inference.
About
Hopsworks is an open-source data-intensive AI platform built around a feature store, founded by the researchers behind Apache Hops. It unifies the ML lifecycle: feature store with online (Redis/MySQL) and offline (Hive/Iceberg) stores, model registry, training pipelines, and serving, plus a vector database for LLM workloads. It positions itself as a sovereign European AI platform for organizations needing data sovereignty and regulatory compliance, and announced in 2026 that Zalando runs its AI infrastructure on it.
The honest caveats: the open-source repository's development pace has slowed (last major activity early 2025), while the commercial AI Lakehouse product gets the new features — check which tier a capability lives in. The Java codebase and full-platform setup are heavyweight for small teams; Feast is lighter if you only need a feature store. Customer announcements are vendor-reported.
What it can do
- Feature store with online and offline tiers
- Model registry with lineage tracking
- Training and feature pipelines
- Model and feature serving endpoints
- Vector database for LLM workloads
Setup
Self-host via the open-source repo or sign up for the managed Hopsworks platform.
Prerequisites
For self-hosting: Kubernetes or bare-metal cluster; data sources to connect.
Use cases
Works with
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 85; Trending score 76/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.