experiment-tracking

What problem does it solve?

This Skill streamlines quantitative research by ensuring experiment tracking workflows are reproducible, have explicit controls, and produce deployable outputs for production trading systems.

Core Features & Use Cases

• Reproducible Research: Define assumptions, equations, and parameters with explicit calibration settings.
• Model Validation: Assess residual structure, numerical stability, and convergence behavior.
• Risk Controls: Implement safeguards for parameter bounds, convergence failures, and model instability.
• Use Case: When deploying a new trading model, use this Skill to validate its performance against defined diagnostics and ensure it meets stability and accuracy limits before release.

Quick Start

Run the experiment tracking diagnostics script on input.csv and save the output to diagnostics.json.