Retrieval Augmented Generation (RAG) with LangChain

Build RAG Systems with LangChain

Retrieval Augmented Generation (RAG) is a technique used to overcome one of the main limitations of large language models (LLMs): their limited knowledge. RAG systems integrate external data from a variety of sources into LLMs. This process of connecting multiple different systems is usually tedious, but LangChain makes this a breeze!

Learn State-of-the-Art Splitting and Retrieval Methods

Level-up your RAG architecture! You'll learn how to load and split code files, including Python and Markdown files to ensure that splits are "aware" of code syntax. You'll split your documents using tokens instead of characters to ensure that your retrieved documents stay within your model's context window. Discover how semantic splitting can help retain context by detecting when the subject in the text shifts and splitting at these points. Finally, learn to evaluate your RAG architecture robustly with LangSmith and Ragas.

Discover the Graph RAG Architecture

Flip your RAG architecture on its head and discover how graph-based, rather than vector-based RAG systems can improve your system's understanding of the entities and relationships in your documents. You'll learn how to convert unstructured text data into graphs using LLMs to do the translation! Then, you'll store these graph documents in a Neo4j graph database and integrate it into a wider RAG system to complete the application.