Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Course Outline
LangGraph and Agent Patterns: A Practical Introduction
- Understanding the differences between graphs and linear chains, and when to use each.
- Exploring agents, tools, and planner-executor loops.
- Creating a 'Hello workflow': establishing a minimal agentic graph.
State, Memory, and Context Passing
- Designing graph state structures and node interfaces.
- Distinguishing between short-term memory and persisted memory.
- Managing context windows, summarization techniques, and data rehydration.
Branching Logic and Control Flow
- Implementing conditional routing and handling multi-path decisions.
- Configuring retries, timeouts, and circuit breaker patterns.
- Managing fallbacks, handling dead-ends, and utilizing recovery nodes.
Tool Use and External Integrations
- Executing function/tool calls from nodes and agents.
- Connecting the graph to external REST APIs and databases.
- Parsing and validating structured outputs.
Retrieval-Augmented Agent Workflows
- Strategies for document ingestion and data chunking.
- Utilizing embeddings and vector stores with ChromaDB.
- Generating grounded responses with citations and protective measures.
Evaluation, Debugging, and Observability
- Tracing execution paths and inspecting node interactions.
- Building golden datasets, conducting evaluations, and running regression tests.
- Monitoring for quality, safety compliance, cost efficiency, and latency.
Packaging and Delivery
- Implementing FastAPI for service deployment and managing dependencies.
- Strategies for graph versioning and system rollback.
- Developing operational playbooks and incident response plans.
Summary and Next Steps
Requirements
- Practical proficiency in Python
- Experience in developing LLM applications or prompt chains
- Familiarity with REST APIs and JSON data formats
Target Audience
- AI engineers
- Product managers
- Developers creating interactive, LLM-driven systems
14 Hours
