Which vector databases does this skill support?

It includes implementation templates for Pinecone, Qdrant, and pgvector (PostgreSQL), covering cloud-native, specialized, and relational options.

Does it support hybrid search?

Yes, the skill includes patterns for combining semantic dense search with keyword or sparse vector search to achieve higher retrieval accuracy.

Is it suitable for large-scale applications?

Absolutely. It covers advanced indexing types like HNSW and IVF+PQ specifically designed for scaling search to millions of vectors with logarithmic latency.

How does this help with RAG systems?

It provides the critical retrieval layer logic, including embedding handling, vector upsertion, and reranking patterns to improve search relevance for LLMs.

Similarity Search Patterns

Name: Similarity Search Patterns
Author: Microck

byMicrock

•

数据科学与机器学习

Implements high-performance similarity search and vector database integration patterns for semantic retrieval.

This skill provides standardized patterns and implementation templates for building efficient similarity search systems using vector databases like Pinecone, Qdrant, and pgvector. It guides developers through selecting appropriate distance metrics, understanding complex index types such as HNSW and IVF, and implementing advanced retrieval features like reranking and hybrid search. Whether you are building a production-grade RAG system or a recommendation engine, this skill ensures your vector search architecture is optimized for both accuracy and latency.

主要功能

01Comprehensive guidance on distance metrics like Cosine, Euclidean, and Dot Product

022 GitHub stars

03Multi-provider templates for Pinecone, Qdrant, and pgvector (PostgreSQL)

04Optimized indexing strategies for different data scales, from Flat to HNSW

05Async implementation patterns for high-concurrency search applications

06Advanced search patterns including cross-encoder reranking and hybrid search

使用场景

01Developing recommendation engines based on content embeddings

02Scaling semantic search to handle millions of vectors with low latency

03Building production-grade RAG (Retrieval-Augmented Generation) systems

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add microck/opendots-microck similarity-search-patterns

For use in Claude.ai and ChatGPT

Download Skill