data-feature-store - Manage ML features from pipelines to serving
Defines, computes, registers, serves, and retrieves ML features while generating point-in-time correct training datasets.
Tags
Updated: 2026-10-05Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Define feature definitions
- Compute batch features
- Compute streaming features
- Serve online features
- Generate training datasets
- Perform point-in-time joins
- Register feature metadata
- Validate and monitor features
- Deploy Feast or Tecton
Inputs
- ML framework
- Inference mode
- Feature sources
- Infrastructure environment
- Online serving requirements
- Feature definition location
Outputs
- Feature store configuration
- Feature definitions
- Serving infrastructure configuration
- Training dataset
- Online feature response
- Feature registry status
- Feature validation status
Requirements
- Feast or Tecton support
- Offline and online stores
- Feature registry access
- Batch processing support
- Streaming processing support
- ML framework integration