RAG & Vector Search
Create private retrieval-augmented pipelines, embedding stores, and local-first vector search systems for secure document intelligence.
Topic breadth
1
Active builds, guides, and subtopic coverage.
Subtopics
RAG Fundamentals
View topicRetrieval-Augmented Generation from first principles: chunking strategies, embedding generation, vector similarity search, context injection, and RAG evaluation with RAGAS.
Vector Databases
View topicSelf-hosted vector databases for sovereign AI memory: pgvector (PostgreSQL extension), Qdrant, ChromaDB, and Milvus. Covers embedding storage, ANN search, and local deployment.
Embedding Models
View topicSovereign embedding model selection and deployment: nomic-embed-text, sentence-transformers, BGE models, and local embedding inference with Ollama — zero data sent to OpenAI.
Advanced RAG
View topicProduction RAG techniques: hybrid search (BM25 + vector), reranking with cross-encoders, multi-hop retrieval, HyDE, query rewriting, and RAGOps monitoring for sovereign pipelines.
Private Document AI
View topicBuild sovereign document Q&A systems: local PDF ingestion, private embedding, self-hosted vector store, and on-device LLM answering — zero documents leave your machine.