Fast Sub-ML Usecase Development using Feature Stores
summary
Argues for Sub-ML: simpler, decision‑support applications that need correctness, transparency, and auditability but not full ML stacks. Uses a Scribble Data feature store plus a feature app store that bundles feature engineering with prebuilt apps for planning, forecasting, pricing, fraud detection, anomaly detection, and metrics. Reports deployment across close to 100 use cases, cutting development to a couple of weeks vs months, lowering coordination and letting one developer build pipelines and consumer apps end‑to‑end.
AI-written summary · about
chapters (5)
- 0:00Introduction
- 1:04Sub-ml use cases explained
- 4:16Traditional vs sub-ml approaches
- 7:05Feature App Store concept
- 9:13Future of sub-ml development
related talks
30mHow Feature Stores Enable Operational ML
17mCreating & operating ML models from event-based data using feature stor...
19mML feature store a cornerstone in modern data architecture for AI/ML
40mCreating and operating ML models from event based data using feature stores
41mHow to Elm-ify Your ML
30mDataEngBytes 2023 - MEL-T2-01 - Mikiko Bazeley
21mTime for quality ML from event-based data
02h 33mHands-On Feature Store and Model Deployment