mle-workflow

What problem does it solve?

This Skill turns ad hoc model work into a production-ready machine learning workflow with explicit contracts, repeatable training, measurable quality gates, and operational safeguards.

Core Features & Use Cases

• Prediction and data contracts: Define inputs, outputs, label timing, freshness rules, and leakage checks before implementation.
• Reproducible training and evaluation: Package configs, baselines, metrics, and artifact tracking so training can be rerun and reviewed reliably.
• Deployment and operations: Prepare serving schemas, rollout criteria, monitoring signals, and rollback paths for forecasting, classification, ranking, and similar ML systems.
• Use case: Review a capacity-exhaustion predictor by checking whether its offline metrics, inference path, and monitoring plan are strong enough to ship safely.

Quick Start

Use the mle-workflow skill to review a production ML pipeline for data contracts, reproducible training, evaluation gates, deployment readiness, and rollback safety.