
Run Thompson Sampling CTR analysis on email treatments. Use when analyzing treatment performance or optimizing email campaigns.
Compare two pipeline versions. Use when validating a new version against baseline.
End-to-end deployment workflow. Runs pipeline, validates, compares, and deploys in one flow.
Create a new pipeline version end-to-end. Use when starting a new feature or major change.
Convert Colab/Jupyter notebooks to production Python scripts. Use when user asks to productionize a notebook, create a Metaflow pipeline from notebook code, or convert prototype to production.
Show current pipeline and deployment status. Use for quick health check.
Run QA validation checks on the recommendation pipeline output. Use after pipeline runs to verify data quality.
Execute BigQuery SQL queries using bq CLI. Use when user asks to run SQL, query data, extract features, or interact with BigQuery tables.
Deploy recommendations to production with full validation. Use after pipeline runs and QA passes.
Debug SQL errors and unexpected results. Use when pipeline fails or produces wrong data.
Implement a feature from a spec file. Use when user says "implement spec", references a spec file to build, or wants to code a planned feature.
Execute the v5.17 vehicle fitment recommendation pipeline. Use when user asks to run the pipeline, refresh recommendations, or generate new recs.
Compare Personalized vs Static treatment performance with unbiased methodology. Use for A/B analysis and treatment comparison.
Generate a team-facing weekly status update from STATUS_LOG.md and git history.
Review code changes for quality, tests, and documentation. Use when user asks to review code, prepare for PR, or check implementation quality.