Which orchestration tools does this skill support?

This skill provides specialized implementation patterns for Kubeflow Pipelines, Apache Airflow, Prefect, and Ray, as well as custom DAG-based workflows.

How does this skill ensure model reproducibility?

The skill enforces strict constraints including pinned dependencies, random state seeding, data versioning via DVC, and the use of containerized training environments.

Is this suitable for both cloud and on-premise ML infrastructure?

Yes, the patterns provided are designed to work across AWS, GCP, Azure, and on-premise Kubernetes clusters using standard MLOps tooling.

Can it help me track experiments and model versions?

Yes, it implements best practices for logging metrics, parameters, and artifacts using tools like MLflow, Weights & Biases, and Neptune to ensure full traceability.

Does it include data validation and quality checks?

Absolutely. It integrates data validation frameworks like Great Expectations to perform schema checks and distribution validation before training begins.

ML Pipeline Engineering

Name: ML Pipeline Engineering
Author: Jeffallan

byJeffallan

•

데이터 과학 및 ML

Orchestrates production-grade machine learning workflows, including feature engineering, automated training, and MLOps lifecycle management.

소개

The ml-pipeline skill transforms Claude into a senior MLOps engineer capable of designing and implementing scalable machine learning infrastructure. It provides expert guidance on building robust feature stores, orchestrating complex training workflows with tools like Kubeflow or Airflow, and managing the complete model lifecycle with experiment tracking and automated deployment. This skill is essential for developers and data scientists who need to move beyond manual scripts to create reproducible, versioned, and observable ML systems that bridge the gap between research and production.

주요 기능

Implementation of production-grade feature stores and transformation pipelines
49 GitHub stars
Orchestration of distributed training workflows using Kubeflow, Airflow, and Prefect
Automated data validation and quality checks using Great Expectations
Experiment tracking and model registry management with MLflow and Weights & Biases
Standardized model versioning and deployment strategies for reproducible AI

사용 사례

Building automated retraining pipelines that trigger based on data drift or performance metrics
Integrating hyperparameter tuning and experiment logging into existing machine learning codebases
Designing a centralized feature engineering system to serve multiple training and inference jobs

소개

주요 기능

Implementation of production-grade feature stores and transformation pipelines
49 GitHub stars
Orchestration of distributed training workflows using Kubeflow, Airflow, and Prefect
Automated data validation and quality checks using Great Expectations
Experiment tracking and model registry management with MLflow and Weights & Biases
Standardized model versioning and deployment strategies for reproducible AI

사용 사례

Building automated retraining pipelines that trigger based on data drift or performance metrics
Integrating hyperparameter tuning and experiment logging into existing machine learning codebases
Designing a centralized feature engineering system to serve multiple training and inference jobs