How does this skill help improve retrieval quality?

It implements advanced techniques like hybrid search (combining dense and sparse vectors), query expansion, and reranking using cross-encoders.

What chunking strategies are available?

It covers four main strategies: Fixed-size for simplicity, Semantic for preserving meaning, Hierarchical for structured docs, and Sliding Window for maximum context.

Does this support local or self-hosted RAG setups?

Yes, it includes configurations for local development using Chroma and self-hosted production setups using Qdrant and Sentence Transformers.

Which vector databases are supported by this skill?

The skill provides implementation patterns and comparison matrices for Qdrant, Pinecone, Chroma, Weaviate, and Milvus.

Can I use this for production-scale AI applications?

Absolutely. These patterns are based on 6+ months of daily production use and cover critical aspects like cost optimization, latency, and scalability.

RAG Implementation Guide

Name: RAG Implementation Guide
Author: applied-artificial-intelligence

byapplied-artificial-intelligence

•

数据科学与机器学习

Implements production-ready Retrieval-Augmented Generation (RAG) systems with optimized vector storage, chunking, and retrieval patterns.

The rag-implementation skill provides a comprehensive framework for building and optimizing Retrieval-Augmented Generation systems. It offers expert guidance on the entire RAG pipeline, including selecting the right vector database (Qdrant, Pinecone, Chroma), implementing advanced chunking strategies like semantic and hierarchical splitting, and integrating high-performance embedding models. It also covers sophisticated retrieval techniques such as hybrid search, query expansion, and reranking to ensure LLMs receive the most relevant context for high-accuracy outputs.

主要功能

01Multi-strategy chunking including fixed-size, semantic, and hierarchical methods

02Vector database configuration for Qdrant, Pinecone, Chroma, and Weaviate

03Advanced retrieval optimization via hybrid search and cross-encoder reranking

04Embedding integration for OpenAI models and local Sentence Transformers

0518 GitHub stars

06Query expansion and metadata filtering patterns for improved recall

使用场景

01Architecting scalable vector database infrastructures for AI applications

02Improving LLM response accuracy by optimizing context retrieval

03Building semantic search engines for large-scale documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add applied-artificial-intelligence/claude-code-toolkit rag-implementation

For use in Claude.ai and ChatGPT

Download Skill

主要功能

01Multi-strategy chunking including fixed-size, semantic, and hierarchical methods

02Vector database configuration for Qdrant, Pinecone, Chroma, and Weaviate

03Advanced retrieval optimization via hybrid search and cross-encoder reranking

04Embedding integration for OpenAI models and local Sentence Transformers

0518 GitHub stars

06Query expansion and metadata filtering patterns for improved recall

使用场景

01Architecting scalable vector database infrastructures for AI applications

02Improving LLM response accuracy by optimizing context retrieval

03Building semantic search engines for large-scale documentation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add applied-artificial-intelligence/claude-code-toolkit rag-implementation

For use in Claude.ai and ChatGPT

Download Skill