feature-eng

What problem does it solve?

ML teams often struggle to produce interpretable, leakage-free features with strong documentation. This Skill provides a disciplined, notebook-driven workflow that enforces domain rationale, leakage checks, and reproducible feature pipelines from audit to deployment.

Core Features & Use Cases

• Auditing existing features for type, distribution, and target relationship to inform safe feature creation.
• Phase-by-phase feature creation with explicit rationale, leakage auditing, and robust documentation in a Jupyter notebook.
• End-to-end pipeline assembly that serializes preprocessing steps and records decisions for reproducibility.

Quick Start

Provide your dataset and target; the system will perform a feature audit, generate defensible features with leakage checks, and output a documented notebook.