Building Dynamic Workflows with LangGraph and LLM Agents Training Cursus

LangGraph and Agent Patterns: A Practical Primer

• Graphs vs. linear chains: when and why
• Agents, tools, and planner-executor loops
• Hello workflow: a minimal agentic graph

State, Memory, and Context Passing

• Designing graph state and node interfaces
• Short-term memory vs. persisted memory
• Context windows, summarization, and rehydration

Branching Logic and Control Flow

• Conditional routing and multi-path decisions
• Retries, timeouts, and circuit breakers
• Fallbacks, dead-ends, and recovery nodes

Tool Use and External Integrations

• Function/tool calling from nodes and agents
• Consuming REST APIs and databases from the graph
• Structured output parsing and validation

Retrieval-Augmented Agent Workflows

• Document ingestion and chunking strategies
• Embeddings and vector stores with ChromaDB
• Grounded responses with citations and safeguards

Evaluation, Debugging, and Observability

• Tracing paths and inspecting node interactions
• Golden sets, evaluations, and regression tests
• Quality, safety, and cost/latency monitoring

Packaging and Delivery

• FastAPI serving and dependency management
• Versioning graphs and rollback strategies
• Operational playbooks and incident response

Summary and Next Steps