LangGraph Foundations: Graph-Based LLM Prompting and Chaining Training Cursus

Introduction to LangGraph and Graph Concepts

• Why graphs for LLM apps: orchestration vs. simple chains
• Nodes, edges, and state in LangGraph
• Hello LangGraph: first runnable graph

State Management and Prompt Chaining

• Designing prompts as graph nodes
• Passing state between nodes and handling outputs
• Memory patterns: short-term vs. persisted context

Branching, Control Flow, and Error Handling

• Conditional routing and multi-path workflows
• Retries, timeouts, and fallback strategies
• Idempotency and safe re-runs

Tools and External Integrations

• Function/tool calling from graph nodes
• Calling REST APIs and services within the graph
• Working with structured outputs

Retrieval-Augmented Workflows

• Document ingestion and chunking basics
• Embeddings and vector stores (e.g., ChromaDB)
• Grounded answering with citations

Testing, Debugging, and Evaluation

• Unit-style tests for nodes and paths
• Tracing and observability
• Quality checks: factuality, safety, and determinism

Packaging and Deployment Fundamentals

• Environment setup and dependency management
• Serving graphs behind APIs
• Versioning workflows and rolling updates

Summary and Next Steps