Hopsworks

Enterprise ML platform with a feature store at its core, providing end-to-end MLOps from feature engineering to model serving.

Visit Website → GitHub

Key Features

✓ Feature groups: Organized feature collections
✓ Time-travel: Point-in-time feature retrieval
✓ Versioning: Feature lineage and versions
✓ Online/offline: Batch and real-time serving
✓ Spark integration: Large-scale processing
✓ Python SDK: Familiar interface
✓ Transformations: Built-in feature transforms
✓ Validation: Data quality checks
✓ Point-in-time joins: Correct historical features
✓ Training datasets: Versioned, reproducible
✓ Data splits: Train/test/validation
✓ Labeling: Target variable management
✓ Model registry: Version and stage models
✓ Online inference: Real-time predictions
✓ Batch inference: Large-scale scoring
✓ Monitoring: Drift detection