Building Dynamic Workflows with LangGraph and LLM Agents Training Course

LangGraph and Agent Patterns: A Practical Introduction

• Graphs vs. linear chains: Identifying the appropriate use case and rationale
• Agents, tools, and the planner-executor loop
• Hello workflow: Constructing a minimal agentic graph

State, Memory, and Context Management

• Defining graph states and node interfaces
• Distinguishing between short-term and persisted memory
• Managing context windows, summarization techniques, and data rehydration

Branching Logic and Control Flow

• Implementing conditional routing and multi-path decision making
• Managing retries, timeouts, and circuit breaker patterns
• Handling fallbacks, dead-ends, and designing recovery nodes

Tool Utilization and External Integrations

• Executing function and tool calls from graph nodes and agents
• Interfacing with REST APIs and databases within the graph structure
• Parsing and validating structured outputs

Retrieval-Augmented Agent Workflows

• Strategies for document ingestion and chunking
• Leveraging embeddings and vector stores using ChromaDB
• Generating grounded responses with citations and safety guardrails

Evaluation, Debugging, and Observability

• Tracing execution paths and analyzing node interactions
• Utilizing golden sets, evaluations, and regression testing
• Monitoring quality, safety, and metrics for cost and latency

Packaging and Deployment

• Serving via FastAPI and managing dependencies
• Version control for graphs and implementing rollback strategies
• Developing operational playbooks and incident response procedures

Recap and Future Directions