Find your next open-source project

Search by name, purpose, or keyword — or paste a GitHub repository URL
All skills

ml-pipeline

Build a complete ML pipeline: Kubeflow or Airflow DAG, MLflow logging, data validation, and deployment config.

When to use it

You have a training script or dataset but no orchestrated, tracked, reproducible pipeline around it, and need stages like validation, training, and model registration wired together.

For example

You say

Use the ml-pipeline skill to turn my scikit-learn churn training script into a Kubeflow pipeline with data validation, MLflow tracking, and a model registry step.

You get
  • Kubeflow pipeline DAG
  • MLflow training script
  • Data validation checkpoint

How it works

  • Data schema and distribution checks run before training and halt the pipeline on failure.
  • Random seeds and dependencies are pinned so runs stay reproducible.
  • Models are only promoted after recorded evaluation thresholds pass.