Welcome to LangChain & LangGraph, the ninth module of the Generative AI & LLM Engineering course. Over the last several modules you built everything by hand: prompt wiring, a RAG pipeline, the agent loop, tool dispatch, memory. That was the right way to learn — you now know exactly what these systems do. But in real projects you rarely rebuild that plumbing from scratch. LangChain and LangGraph are the most widely used frameworks for giving you those pieces as reusable building blocks, so you can focus on your application instead of the loop.

You’ll start with why frameworks exist and when they help (and when raw SDK code is still better). Then you’ll learn LangChain basics — chat models, prompt templates, and the LCEL “pipe” syntax that composes components into a chain. You’ll rebuild your RAG pipeline with LangChain in a fraction of the code, using its loaders, splitters, vector stores, and retrievers. You’ll move to LangGraph to build agents — the same think-act-observe loop from Module 8, now provided for you — and add memory with a checkpointer. The capstone is a LangGraph agent that combines a retrieval tool, a calculator, and memory to answer real multi-step questions.

Every example runs for real on LangChain 1.x and LangGraph 1.x against the Claude API using the affordable claude-haiku-4-5 model, with embeddings handled locally and freely. Because you already built these systems by hand, you’ll recognize exactly what the framework is doing under the hood — which is the best possible position from which to use one well.

Start with Lesson 1, where you’ll learn what frameworks give you and when to reach for one.