Can I use this for experiment tracking?

Absolutely. It integrates with experiment tracking tools like MLflow and Weights & Biases to log hyperparameters, metrics, and artifacts.

How does it handle data versioning?

The workflow includes strategies for data versioning and lineage tracking to ensure every model training run is fully reproducible.

Does it support model monitoring in production?

Yes, it includes patterns for configuring performance monitoring, drift detection, and automated rollback mechanisms for deployed models.

Which orchestration tools does this skill support?

This skill provides patterns and guidance for major orchestration platforms including Apache Airflow, Kubeflow Pipelines, Dagster, and Prefect.

ML Pipeline Workflow

Name: ML Pipeline Workflow
Author: HermeticOrmus

byHermeticOrmus

データサイエンスとML

Orchestrates end-to-end MLOps pipelines from data preparation and training to validation and production deployment.

概要

This skill provides a comprehensive framework for building production-grade MLOps pipelines, covering the entire machine learning lifecycle from data ingestion to monitoring. It offers architectural guidance for DAG-based orchestration, robust data validation, and automated deployment strategies like canary and blue-green releases. By using this skill, developers can ensure reproducibility, modularity, and scalability in their machine learning workflows while integrating seamlessly with industry-standard tools like Airflow, Kubeflow, and MLflow.

主な機能

End-to-end DAG orchestration design patterns
Automated data validation and feature engineering workflows
Integration for model training and experiment tracking
Standardized model validation and A/B testing frameworks
Automated deployment patterns for production environments
0 GitHub stars

ユースケース

Building reproducible ML training pipelines from scratch
Designing modular orchestration for production ML systems using Airflow or Kubeflow
Implementing automated data-to-model deployment CI/CD workflows

概要

主な機能

End-to-end DAG orchestration design patterns
Automated data validation and feature engineering workflows
Integration for model training and experiment tracking
Standardized model validation and A/B testing frameworks
Automated deployment patterns for production environments
0 GitHub stars

ユースケース

Building reproducible ML training pipelines from scratch
Designing modular orchestration for production ML systems using Airflow or Kubeflow
Implementing automated data-to-model deployment CI/CD workflows