ML Model Governance Cheat Sheet

Versioning, approval, and drift tracking

Last Updated: November 21, 2025

Focus Areas

Focus
Version models and log approvals before deployment
Monitor drift metrics and retrain when thresholds exceed limits

Commands & Queries

mlflow models list
List registered models
mlflow models stage-transition --name model --stage Staging
Promote a model
python scripts/check_drift.py
Run drift analysis

Summary

Treat models like products with documentation, approvals, and drift guardrails.

💡 Pro Tip: Record governance artifacts in a single registry and refresh approvals quarterly.
← Back to Data Science & ML | Browse all categories | View all cheat sheets