LangGraph: From Basics to Advanced AI Agents with LLMs

Name: LangGraph: From Basics to Advanced AI Agents with LLMs
Rating: 4.6 (48 reviews)

Learn how to build custom AI Agents with LangGraph

Created byTensor Teach

Last updated 5/2025

English

English [Auto],

What you'll learn

Software Engineers looking to build Generative AI applications with LangGraph.
Data Scientists looking to build Generative AI applications with LangGraph.
Learners looking to further their knowledge beyond LangChain.
Generative AI enthusiasts looking to learn the latest methods for building AI agents.

Course content

3 sections • 19 lectures • 1h 38m total length

What Is LangGraph?3:27
Explore LangGraph, a python package for building agentic workflows as graphs with nodes, edges, and cycles, integrating with Lang Chain to track autonomous agents.
Drafting Our First Agent5:11
Draft your first LangGraph agent graph by sketching a manager routing tasks to two nodes, using conditional edges and coin-flip logic.
Creating The Agents State4:50
Translate the flow into a graph state by defining a typed dict with input, assigned node, and prev node to write, then propagate the state through manager and worker nodes.
Building Our Nodes5:09
Create and wire nodes as Python functions that update a state, using a bad manager with a coin flip to assign the next node and update prev and assigned fields.
Note on Conditional Edges & Routing Functions0:18
Building Our Edges6:37
Define and connect conditional edges in a LangGraph workflow, using routing functions to move state among the bad manager and worker nodes to node one, node two, or the end.
Compiling Our Graph & Testing It9:09
Compile and test a graph by defining nodes and conditional edges, wiring a shared state, compiling the graph, visualizing with mermaid, and running tests with a human input.
Section Notebook - LangGraph Core Concepts0:03

Outlining Our News Writer Agent7:21
Design a news writer agent in LangGraph that takes a query, uses a search tool to fetch articles, outlines the article with an outliner, and writes the final article.
Defining State & Tools4:03
Define a graph state with a messages list, load the Tivoli search tool, and configure OpenAI and Tivoli keys to power LangChain agents that generate outlines.
Defining Nodes6:47
Define and orchestrate agents in a lang chain system by wiring system messages, tools, and state to create searcher, outliner, and writer nodes, then assemble tool nodes and edges.
Adding Edges & Compiling Our News Writer Graph6:42
Add edges and compile the LangGraph News Writer by routing conditional edges from the search node to tools or outliner, then connect tool, outliner, and writer.
Testing Our News Writer Agent6:01
Explore testing a news writer agent within a LangChain workflow, including input messages, tool calls, and an outliner. The lesson highlights editor use, human in the loop, and reflection.
Section Notebook0:03

Adding Reflection to Our News Writer Agent10:20
Enhance your news writer agent with reflection by routing its output through an editor node that critiques and revises the article until it meets quality guidelines.
Human-In-The-Loop9:20
Replace the AI editor with a human in the loop by saving workflow state with a memory saver (check pointer), interrupting before editing, and delivering targeted feedback via a thread.
Building Custom Tools in LangGraph9:33
Build a custom publish tool in LangGraph by adding a publisher node and a tool node with the tool decorator, converting articles to json payloads for api publishing.
Section Notebook0:03
Bonus Lecture: Free AI Research Newsletter Offering3:18
Discover how the Behind the Hood AI newsletter curates weekly arXiv AI papers, highlighting multimodal llms, embodied agents, robotics, and quantization, with concise research questions, methods, results, and takeaways.

Requirements

Some python programming experience
Previous experience with LangChain

Description

Embark on a comprehensive journey into the world of AI agents with LangGraph. This course is designed to guide you from fundamental concepts to advanced techniques, equipping you with the skills to build sophisticated AI systems. Starting with the core principles, you'll learn about graphs, nodes, edges, and states, and see how they form the foundation of LangGraph workflows. The course begins with constructing a basic agent, allowing you to grasp the essentials through hands-on practice.

Next, you'll dive deeper by building a News Writer Agent, enhancing your understanding by integrating state and tools into your agents. The focus will be on practical applications, ensuring you can visualize and test your agents effectively. Finally, the course introduces advanced techniques, including reflection, human-in-the-loop processes, checkpointers, and threads. You'll also learn to incorporate custom tools, adding versatility and functionality to your agents. Whether you're a beginner or looking to advance your skills, this course provides a structured, step-by-step approach to mastering AI agent development with LangGraph.

The goal of this course is to equip you with the understanding and skills you need to build your own agents. There are plenty of off-the-shelf agents available via LangGraph and other resources. However, in our experience, when building agents for production you will need to be able to customize. At the end of this course, it is our goal to make sure that you are capable of building your own custom workflows in LangGraph.

Note: Prior python programming experience and some experience with LangChain are required for this course.

Who this course is for:

Developers, Data Scientists, AI Engineers, and Enthusiasts looking to learn how to build AI Agents in LangGraph

LangGraph: From Basics to Advanced AI Agents with LLMs

What you'll learn

Explore related topics

Course content

LangGraph Core Concepts8 lectures • 35min

Building a News Writer Agent6 lectures • 31min

Adding Advanced Capabilities to our Agent5 lectures • 33min

Requirements

Description

Who this course is for: