# Ml Experiment Tracking Pro

> Rigorous tracking of parameters, metrics, and artifacts for reproducible machine learning.

- Skill: `jcorpac/ml-experiment-tracking-pro` (Agent Skill)
- Install (CLI): `npx skillmds@latest add jcorpac/ml-experiment-tracking-pro`
- Raw SKILL.md: https://api.skillmd.com/api/skills/jcorpac/ml-experiment-tracking-pro/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: jcorpac (https://skillmd.com/u/jcorpac)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/jcorpac/ml-experiment-tracking-pro

---


# ML Experiment Tracking Pro

ML development is non-linear. Tracking every variant is essential for finding the "winning" model.

## Core Concepts
- **Runs**: A single execution of a training script.
- **Experiments**: A collection of related runs (e.g., "Optimizing Learning Rate").
- **Artifacts**: Storing models, plots, and datasets.

## Frameworks
- **MLflow**: Open source, platform-agnostic tracking server.
- **Weights & Biases (W&B)**: Collaborative, visual platform for deep learning teams.

## Best Practices
- **Auto-Logging**: Use `mlflow.autolog()` for effortless tracking in supported frameworks.
- **Tagging**: Use tags to mark "Production Ready" or "Baseline" models.
- **Nesting**: Use nested runs for hyperparameter tuning trials.