NirDiamant/RAG_Techniques ? reverse-engineered prompt

Build me a public learning repo for advanced retrieval augmented generation techniques. I want it to feel like a practical tutorial library for people who already know the basics and want better answers from their AI search systems.

Make each technique a clear notebook with plain English explanations, simple diagrams or images where helpful, sample data, runnable code, and a short section showing what problem it solves. Include reusable helper functions so the notebooks don’t repeat too much setup. Also add runnable script versions for people who don’t want to use notebooks.

Please cover ideas like better retrieval, better chunking, reranking, query rewriting, evaluation, agentic RAG, and comparing results so users can see when a technique helps or fails. Use common Python AI tools and keep API keys in environment variables.

Add a good README that explains the purpose of the repo, how to install and run things, and how people can contribute. Look up current docs online if you need to.