nirdiamant/rag_techniques

This repository is a curated collection of advanced techniques for building Retrieval-Augmented Generation (RAG) systems, with each technique presented as its own tutorial notebook. RAG is the approach where an AI assistant first looks up relevant information from a body of documents and then uses a large language model to write an answer grounded in what it found. The goal is to make responses more accurate, more relevant to the source material, and less prone to inventing facts. The problem the repo addresses is that a basic RAG setup, chunk some documents, find the most similar chunks, paste them into a prompt, often works poorly in practice. Answers can miss the point, leave out useful context, or pull in noisy material. The repository walks through more advanced techniques aimed at improving accuracy, efficiency, and contextual richness, so practitioners can move past the naive baseline. Each technique ships as a detailed tutorial notebook, so you can read the explanation, run the code, and adapt it to your own project. The repo is positioned as a hub for researchers and practitioners experimenting with RAG, and the description mentions integrations with common AI building blocks for working with language models and embeddings. You would use this when building a search-aware AI assistant, a question-answering tool over a private document set, or any application where an LLM needs to consult external knowledge before responding. The notebooks are written in Python and presented as Jupyter notebooks. The full README is longer than what was provided.