What types of ML problems can this skill design for?

It is designed for a wide range of problems including recommendation systems, fraud detection, search ranking, content moderation, and customer churn prediction.

Does this skill help with model monitoring?

Yes, the generated designs include components for monitoring, observability, and preventing training-serving skew.

Can it handle LLM and RAG architectures?

Absolutely. The skill includes specific workflows for LLM-based systems, RAG architecture, and vector database selection.

Does it provide specific technology recommendations?

Yes, the output includes a recommended technology stack for ingestion, processing, training, and inference based on your specific scale and latency needs.

ML Pipeline Designer

Name: ML Pipeline Designer
Author: melodic-software

bymelodic-software

•

Data Science & ML

Designs end-to-end machine learning system architectures tailored to specific business requirements and scalability needs.

The ML Pipeline skill empowers developers to architect robust, production-grade machine learning systems by automating the design of data ingestion, feature engineering, training, and deployment workflows. It leverages specialized subagents to address complex domain-specific challenges such as latency constraints, training-serving skew, and model observability. Whether you are building a real-time fraud detection engine or a large-scale recommendation system, this skill provides comprehensive architectural blueprints, technology stack recommendations, and phased implementation plans to ensure project success.

Key Features

01Automated analysis of latency (real-time vs. batch) and scalability requirements

0238 GitHub stars

03Strategies for mitigating training-serving skew and model versioning

04End-to-end architecture design for data processing, feature stores, and model serving

05Specialized subagent support for RAG, LLM patterns, and vector databases

06Comprehensive tech stack recommendations including monitoring and A/B testing

Use Cases

01Building multi-stage search ranking systems with vector search integration

02Architecting a high-throughput recommendation system for millions of active users

03Designing low-latency fraud detection pipelines for real-time payment processing

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add melodic-software/claude-code-plugins ml-pipeline

For use in Claude.ai and ChatGPT

Download Skill