📄️ LangChain
Build a real assistant with LangChain: chains, retrievers, memory, tools and agents. 6 hours, applied projects, a 40-question exam and a verifiable certificate.
📄️ 1. What LangChain adds
Module 1 of the LangChain premium course: the abstractions that unify providers, the cost of that abstraction, when to write direct calls, and the state of the ecosystem.
📄️ 2. Models, prompts, parsers
Module 2 of the LangChain premium course: chat model interfaces, prompt templates, structured output with Pydantic, native structured outputs and handling parse failures.
📄️ 3. Composing chains
Module 3 of the LangChain premium course: the pipe operator, RunnablePassthrough and RunnableParallel, conditional branching, streaming and batching a chain.
📄️ 4. Loaders and splitting
Module 4 of the LangChain premium course: PDF, web and office loaders, recursive character splitter, chunk metadata and how the running expense assistant ingests receipts.
📄️ 5. Vector stores and retrievers
Module 5 of the LangChain premium course: the VectorStore interface, Chroma and FAISS, similarity threshold, multi-query retrieval and a minimal RAG chain on the expense policy.
📄️ 6. Conversation memory
Module 6 of the LangChain premium course: message history, window and summary memory, per-session storage, persistence and controlling context window growth.
📄️ 7. Tools and function calling
Module 7 of the LangChain premium course: declaring a tool from a typed function, native tool calling, argument validation, and the tools of the running expense assistant.
📄️ 8. Agents
Module 8 of the LangChain premium course: tool-calling agents, LangGraph for explicit control flow, iteration limits, and human-in-the-loop checkpoints.
📄️ 9. Tracing, evaluation, debug
Module 9 of the LangChain premium course: execution traces, cost per call, evaluation datasets, evaluators, and regression testing between prompt and model versions.
📄️ 10. End-to-end project
Module 10 of the LangChain premium course: assembling the expense assistant end to end, minimal interface, tool-error handling, tests, and going to production.
📄️ Recap and exam
Complete recap of the LangChain premium course: chains, retrievers, memory, tools and agents, then the 40-question exam and the verifiable certificate.