ML Experiment Tracker

This skill provides guidance for systematic machine learning experimentation with proper tracking, versioning, and reproducibility practices.

Core Competencies

• Experiment Tracking: MLflow, Weights & Biases (wandb), Neptune, Comet
• Data Versioning: DVC, Delta Lake, LakeFS
• Model Registry: Version control for trained models
• Reproducibility: Environment, code, data, and hyperparameter tracking

Experiment Tracking Fundamentals

What to Track

Every experiment should log: