LangGraph AI Agent Development
This course takes developers from LangGraph fundamentals to production-grade AI agents. Learn graph-based state management, LLM integration, tools, conditional workflows, persistence, RAG, multi-agent systems, human approval, MCP, FastAPI, LangSmith observability, security, durable execution, and scalable architecture.
Curriculum Areas
- Module 1: LangGraph Fundamentals
- Module 2: Python for LangGraph
- Module 3: LangGraph Core Concepts
- Module 4: Building AI Agents
- Module 5: LLM Integration
- Module 6: Tools & Function Calling
- Module 7: Conditional Workflows
- Module 8: Memory & Persistence
- Module 9: RAG with LangGraph
- Module 10: Multi-Agent Systems
- Module 11: Human-in-the-Loop
- Module 12: LangGraph + MCP
- Module 13: LangGraph Production Development
- Module 14: LangSmith & Observability
- Module 15: Agent Security
- Module 16: Advanced LangGraph
LangGraph AI Agent Curriculum
- What is LangGraph?
- LangChain vs LangGraph
- Graph-based AI application architecture
- Nodes, edges and states
- Graph execution model
- State management
- Building your first LangGraph application
- Python fundamentals
- Functions and classes
- Type hints
- Async programming
- Pydantic
- Environment variables
- API integration
- Virtual environments
- StateGraph
- State schemas
- Nodes
- Edges
- Conditional edges
- Entry and exit points
- Graph compilation
- Graph execution
- State updates
- What are AI Agents?
- Agent architecture
- LLM-powered agents
- Tool calling
- Function calling
- Agent decision-making
- Agent loops
- Human-in-the-loop workflows
- OpenAI models
- Anthropic Claude
- Google Gemini
- Azure OpenAI
- Local/open-source LLMs
- Model configuration
- Streaming responses
- Structured output
- Creating custom tools
- Tool schemas
- Calling APIs from agents
- Database tools
- Search tools
- Calculator and utility tools
- Tool error handling
- Dynamic tool selection
- Conditional routing
- Agent decision nodes
- Branching workflows
- Loops
- Retry mechanisms
- Error handling
- Multi-step reasoning workflows
- Short-term memory
- Long-term memory
- Checkpoints
- Conversation history
- Persistent state
- Thread management
- State recovery
- RAG architecture
- Document loading
- Text splitting
- Embeddings
- Vector databases
- Retrieval
- Context generation
- Building a RAG Agent
- Advanced RAG workflows
- Multi-agent architecture
- Agent-to-agent communication
- Supervisor agents
- Specialized agents
- Parallel agents
- Agent handoffs
- Collaborative AI agents
- Multi-agent workflow design
- Human approval workflows
- Interrupts
- Agent review
- Manual intervention
- Approval/rejection flows
- Resume execution
- Secure agent actions
- Model Context Protocol fundamentals
- MCP servers
- MCP tools
- Connecting agents to MCP
- External tool integration
- Secure MCP architecture
- Building MCP-enabled agents
- API development with FastAPI
- Streaming
- Authentication
- Logging
- Error handling
- Rate limiting
- Environment management
- Production architecture
- LangSmith fundamentals
- Tracing
- Debugging
- Agent monitoring
- Prompt evaluation
- Token usage
- Latency monitoring
- Evaluation datasets
- Production observability
- Prompt injection
- Tool abuse
- Data leakage
- Agent permissions
- Secrets management
- Input/output validation
- Secure tool execution
- AI Agent security best practices
- Subgraphs
- Parallel execution
- Dynamic workflows
- Durable execution
- Advanced state management
- Fault tolerance
- Complex agent orchestration
- Scalable agent architecture