Module Overview
This module covers LangGraph as the framework for stateful agent workflows. It progresses from core graph concepts and state management to tool-calling agents, human-in-the-loop control, persistence, multi-agent systems, and deployment on the LangGraph Platform.
Learning Objectives
- Explain LangGraph's graph-based execution model and core concepts.
- Manage state and define nodes, edges, and conditional routing.
- Implement the ReAct pattern and human-in-the-loop checkpoints.
- Use memory and persistence to resume sessions.
- Orchestrate multi-agent systems and add streaming and observability.
Topics Covered
LangGraph
- Foundations and setup
- Core graph concepts
- State management
- Nodes, edges, and routing
- Tool-calling and ReAct
- Human-in-the-loop
- Memory and persistence
- Multi-agent systems
- Streaming and observability
- The LangGraph Platform
Key Concepts & Terminology
State graph, reducers, conditional edges, ReAct loop, checkpointer, interrupt/resume, supervisor and worker agents, streaming modes.
Tools & Frameworks Referenced
LangGraph, LangGraph Platform, LangSmith for tracing.
Prerequisites
Module 19 (function calling and structured outputs).