ML Workflow

Design machine learning workflows covering experiment tracking, training, serving, and drift monitoring.

1|1|Updated Feb 21, 2026

Search Tags:#workflow #drift detection #mlops #machine learning #model training #ml #model serving

Ready to run

npx skills add https://github.com/dtsong/claude-code-windows-setup --skill ml-workflow

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ML Workflow
Source: https://github.com/dtsong/claude-code-windows-setup/tree/main/skills/council/alchemist/ml-workflow
Command: npx skills add https://github.com/dtsong/claude-code-windows-setup --skill ml-workflow

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill streamlines the complex process of designing and implementing machine learning workflows, from initial problem definition to ongoing monitoring.

Core Features & Use Cases

End-to-End Design: Covers feature engineering, experiment tracking, training pipelines, model serving, A/B testing, and drift monitoring.
Tool Agnostic: Provides a framework for selecting and integrating various ML tools and platforms.
Use Case: A data science team needs to build a new recommendation engine. This Skill helps them define the problem, select features, choose an experiment tracking tool like MLflow, design the training and serving infrastructure, and set up monitoring for performance degradation.

Quick Start

Design an ML workflow for a classification problem using the ML Workflow skill.

Frequently Asked Questions about ML Workflow

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I design an end-to-end machine learning workflow for model training and serving?▼

Designing an end-to-end machine learning workflow involves defining pipelines for feature engineering, model training, serving, and drift monitoring. This process ensures reproducible research and scalable model deployment across diverse infrastructure constraints.

What is drift detection in MLOps and when do I need it?▼

Drift detection in MLOps monitors performance degradation in production models. You need it as part of your ML workflow to track data and model behavior changes over time, ensuring continuous model improvement and robust serving.

How do I set up experiment tracking and feature stores for a new recommendation engine?▼

Setting up experiment tracking and feature stores requires selecting tool-agnostic ML platforms to manage features and log experiments. This framework integrates with your training pipelines to maintain reproducible research and efficient MLOps.

Can I build an ML workflow pipeline that supports diverse data sources and infrastructure constraints?▼

Yes, you can build an ML workflow pipeline supporting diverse data sources and infrastructure constraints. The framework provides tool-agnostic design selections to address diverse ML problem types for scalable and robust model deployment.

What's the best way to structure ML pipelines for continuous model improvement and A/B testing?▼

Structuring ML pipelines for continuous improvement involves integrating A/B testing, model serving, and drift monitoring into the training workflow. This approach addresses performance degradation and satisfies requirements for efficient MLOps.

Does this ML workflow approach work for classification problems and other diverse ML problem types?▼

Yes, this ML workflow approach works for classification problems and other diverse ML problem types. It provides a comprehensive framework encompassing experiment tracking, feature stores, and serving tailored to your specific data sources.