LangGraph Introduction
Overview
LangGraph Introduction is a two-day, hands-on course on building reliable agentic applications with open source LangGraph and related LangChain tools. LangGraph provides a graph-based runtime for stateful agent workflows, letting developers control how models, tools, memory, human review, and application state interact over time. Participants see how LangGraph fits into the wider AI development landscape and how it relates to LangChain components. They also learn to debug, test, observe, and serve LangGraph applications without relying on hosted platforms.
The course moves from graph fundamentals through tool-using agents, branching workflows, persistence, human-in-the-loop patterns, local observability, evaluation, and self-hosted deployment. Labs build cumulatively. Attendees start by scaffolding a project and running a first graph workflow. They then add explicit state transitions to build a branching application with nodes, edges, and conditional routing. From there they wire in tool calling and structured outputs. They then add checkpointing and a human approval step so a workflow can pause for review and resume where it left off.
Day two shifts into agent design, where students build a supervisor-style workflow that coordinates tool-using workers. They then trace, evaluate, and improve that application with local, open source tooling rather than a hosted platform. The final labs package the application behind an API-style interface. Attendees run a production readiness review that checks guardrails, observability, and failure recovery before the workflow ships.
Duration
2 Days
Who Should Take This Course
Prerequisites
Participants should have intermediate programming skills, preferably in Python, and basic familiarity with LLM applications, prompting, APIs, and command-line development. Experience with LangChain, agentic systems, or cloud deployment is helpful but not required.
Why You Should Take This Course
Upon completing LangGraph Introduction, attendees will be able to:
- Explain how LangGraph supports stateful, controllable, long-running agent workflows
- Build graph-based LLM applications using nodes, edges, state, tools, and conditional routing
- Add persistence, checkpointing, human-in-the-loop controls, streaming, and memory to graph workflows
- Integrate LangGraph with LangChain components, model providers, tools, retrievers, and structured outputs
- Debug, trace, evaluate, and improve LangGraph applications with open source techniques and local tooling
- Package and serve LangGraph applications using standalone, containerized, and self-hosted options Apply reliability, security, testing, and operational practices for production agent systems
Course Outline
LangGraph Introduction
- LangGraph and the LangChain Ecosystem
- Graph Fundamentals: State, Nodes, and Edges
- Models, Tools, and Structured Outputs
- Persistence, Memory, and Human-in-the-Loop Workflows
- Agent and Multi-Agent Patterns with LangGraph
- Tracing, Debugging, and Evaluation
- Packaging, APIs, and Self-Hosted Deployment
- Reliability, Security, and Production Readiness