Skip to content

Chapter 1.7 — Experiment tracking and model registry

🎯 Objective

Treat experiments as engineering artifacts.

🧠 Concept

  • Experiment tracking — recording parameters, code, data, metrics, artifacts and environment.
  • Model registry — a versioned catalog of models with metadata, status (staging, production, archived) and approval.

Tools: MLflow, Weights & Biases, DVC, ClearML.

🚨 Failure modes

  • A "laptop model" with no traceability.
  • A metric reported in a deck differs from the reproducible metric.
  • A model in production with no traceable dataset/code.

📝 Mandatory templates

  • Model card — purpose, dataset, metrics, limits, risks.
  • Dataset card — origin, schema, bias, restrictions.
  • Eval report — versioned comparison between the current model and the candidate.
  • EX-ML-03 — scikit-learn pipeline + MLflow + local model registry.

📌 Checklist

  • [ ] Does every experiment have an ID and traceability?
  • [ ] Does the production model have a published model card?
  • [ ] Does promotion between staging and production require approval?

📚 References