⛓️Gen AI
LangChain
Build production LLM apps with LangChain
Learn LangChain from scratch — chains, prompts, memory, RAG pipelines, tools, and agents. Build real AI applications powered by OpenAI and HuggingFace models.
36 lessons
08 hrs
Advanced
What you will learn
Build real-world projects from scratch
Write clean, production-ready code
Understand core concepts deeply with diagrams
Follow industry best practices
Get hands-on with code in every lesson
Access lifetime updates as the tech evolves
Curriculum
Module 1Chat Models and Prompt Templates2 lessons
Module 2Output Parsers — Structured Output, Not Just Raw Text2 lessons
String and JSON Output Parsers — Extracting Usable Data12 min
Pydantic-Validated Structured Output — Guaranteed Schema, Not Just Valid JSON13 min
Module 3The Runnable Protocol and LCEL's Pipe Operator2 lessons
The Pipe Operator — Chaining Steps Without Manual Wiring13 min
The Uniform Runnable Interface — invoke, stream, and batch13 min
Module 4Streaming and Async — Why LCEL Provides Both for Free2 lessons
Sync vs Async — The Real Concurrency Difference, Measured13 min
Streaming a Real LangChain Response Through FastAPI14 min
Module 5RunnableParallel — Running Independent Steps Concurrently2 lessons
RunnableParallel — Multiple Independent Chains, One Input13 min
Measuring RunnableParallel's Real Speed Advantage12 min
Module 6RunnableBranch and Routing — Conditional Logic Without If/Else Sprawl2 lessons
RunnableBranch — Routing to Different Chains Based on a Condition13 min
Building a Real, LLM-Classified Router13 min
Module 7Building a Complete Multi-Step Chain1 lesson
Assembling Sequential, Parallel, and Branching Logic Into One System15 min
Module 8Conversation Memory — Maintaining Context Across Turns2 lessons
Why a Plain Chain Forgets Everything Between Calls12 min
Automatic Message History With RunnableWithMessageHistory14 min
Module 9Managing Long Conversations — Windowing and Summarization2 lessons
Measuring the Real Cost of Unbounded Conversation History13 min
Windowing and Summarization — Two Real Fixes, Measured14 min
Module 10Document Loaders and Text Splitters — Getting Real Documents Ready2 lessons
Document Loaders — From Real Files to Usable Text12 min
Text Splitters — Why Naive Splitting Breaks Meaning13 min
Module 11Embeddings and Vector Stores — Making Chunks Genuinely Searchable2 lessons
LangChain's Embedding Model Interface — Standardized, Like Chat Models12 min
Building and Querying a Real Vector Store14 min
Module 12Building Retrievers — The Retriever Abstraction and Diverse Results2 lessons
The Retriever Abstraction — A Vector Store as a Runnable12 min
Maximal Marginal Relevance — Fixing Redundant Retrieval Results13 min
Module 13Building a Complete RAG Chain2 lessons
Assembling Retrieval and Generation Into One Chain14 min
Proving RAG Actually Changes the Answer13 min
Module 14Evaluating RAG Quality2 lessons
Measuring Retrieval Quality — Did It Find the Right Documents?13 min
Detecting When Generation Drifts From Retrieved Context14 min
Module 15Tool Calling and Function Binding2 lessons
Binding Real Functions as Tools the Model Can Call13 min
Building a Complete Tool-Calling Loop14 min
Module 16Building a Simple Tool-Using Agent2 lessons
Why Module 15's Fixed Loop Breaks on Multi-Step Questions12 min
Building a Genuine Agent With create_agent14 min
Module 17Agent Limits and When You Need LangGraph Instead2 lessons
Constructing a Task This Course's Simple Agent Cannot Handle Well13 min
What This Course Covers, and What LangGraph Solves10 min
Module 18Error Handling, Retries, and Basic Observability2 lessons
Handling Real Failures With Retries and Exponential Backoff13 min
Basic Observability — Seeing What Actually Happened12 min
Module 19Deploying a LangChain Application1 lesson
Assembling a Complete, Production-Hardened Application15 min
36 total lessons across 19 modulesPreview Module 1 free