Navigating the Skies of Data: A Flight Through Feature Stores
summary
Explains what a feature store does, shows a live demo creating a wine feature table in Databricks Feature Store (create_table call, dataframe + primary key) and verifies it appears in the UI, and highlights automated lineage between feature table versions and model training for compliance. Compares offline lakehouse storage with online serving via Redis/Cosmos for sub‑second inference, names alternatives (Feast, Hopsworks, SageMaker, Azure ML), and warns about cost, governance and data quality risks.
AI-written summary · about
chapters (13)
- 0:00Feature stores journey
- 1:15What is a feature
- 3:23Feature vs fact tables
- 5:11Feature reuse and discovery
- 7:41Managed vs self-hosted
- 8:03Demo: Azure Databricks feature store
- 10:12Registering feature tables
- 11:58Model lineage and compliance
- 13:02Core benefits
- 15:08Team collaboration
- 16:32Data quality best practices
- 17:58When to adopt
- 19:20Q&A
related talks
08mFeature Store Usage Patterns - From a Single Data Scientist to an Enterprise
35mPanel 03- How to make the Feature Store become a hidden part of the productionalization of ML models
30mHow Feature Stores Enable Operational ML
10mEmpowering your feature stores with AI feature discovery
32mA feature Store for Machine Learning
33mFeathr, an Open Source, Battletested, Scalable and Enterprise-Grade Feature Store
30mAmazon SageMaker Feature Store: Store, discover, & share features for ML apps
51mMetaConfig driven FeatureStore