Your best Bet: Effortless MLOps with Python Models in
summary
Claims that dbt 1.3 plus Python models run on Databricks can implement a full ML pipeline: data prep in SQL, feature snapshots, experiments (cross‑validation, hyperparameter tuning), and daily inference. The demo uses the European Soccer Dataset and a packaged library risk_rover, trains simple models, and records about €25 total profit on the test slice (~1.5 euro‑cents average per match), so the pipeline works but yields only small, fragile gains and requires careful snapshots, surrogate keys, and packaging.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:00Agenda
- 1:50Why Python
- 3:13What is dbt
- 6:42How to use dbt
- 8:19The Bookies Data Warehouse
- 14:27Demo
- 19:22Design Patterns
- 26:09Conclusion
related talks
51mHow to get started with MLOps in Azure Databricks
13mMLOps at Scale
16mKeynote: DBRX: Training Modern LLMs on Databricks
48mMachine Learning in Azure Databricks
26mMLOps for the rest of us- A poor man's guide to putting models in production
32mMLOps, governance and trust
45mMLOps and ML-Flops: going from dev to prod
32mMaking Models Work: How an optimized ML lifecycle can drive better ROI