Watching lessons is only half of learning. These five projects take you from a first document chatbot to a deployed, tested RAG app. Each project comes with a video that walks through a build, so you can code along and then adapt it to your own documents.
They are ordered from easiest to hardest, and each one draws on modules from the course. You do not need to do them all, but finishing two or three of them and putting the code on GitHub makes a strong portfolio.
Build a question-answering app with LangChain, Chroma and an OpenAI model that loads text documents, stores chunk embeddings and answers from them. Then swap in your own documents. Uses modules 4 to 8.
Video 16minMake a RAG app that answers questions about PDFs and runs on your own computer with Ollama. It also updates the database when files change and unit tests the quality of its answers. Uses modules 6 to 8 and 11.
Video 21minBuild an adaptive, self-reflective RAG app with LangGraph, Groq, Chroma and Tavily web search, where the system checks its own results and corrects itself. Uses modules 9 and 10.
Video 13minAdd an evaluation pipeline to one of your projects: find failures in traces, collect them in a dataset, score them with automated judges, and test each change with an experiment. Uses module 11.
Video 6minTake a LangChain PDF chat app from local to live, with the backend on DigitalOcean, the frontend on App Platform and a managed Postgres database. Uses module 12.
Video 36min