Why Graph Orchestration Exists
By the end of this module you can say what a graph gives you that a chain does not - branching, cycles and resumability - and recognise when a chain was enough.
LangGraph for Agentic Workflows
An approval-aware LangGraph workflow with state, deterministic steps, model-driven steps, checkpointing, resume behavior, and trace evidence.
State, Nodes, Edges, START, and END
By the end of this module you can define state, nodes and edges, and describe how a value moves through a graph from START to END.
Deterministic Steps versus LLM Steps
By the end of this module you can decide which nodes need a language model and which are ordinary functions, and stop paying for judgement where logic will do.
Conditional Routing and Error Paths
By the end of this module you can route conditionally between nodes and give the graph a real error path rather than letting a failure fall through it.
Tool Nodes and Controlled Action Execution
By the end of this module you can execute tools inside a graph with controlled arguments, so an action happens only in the state where it is valid.
Persistence and Checkpointing
By the end of this module you can checkpoint a workflow so a long run survives a restart, and resume from where it stopped rather than from the beginning.
Human-in-the-Loop Review and Edits
By the end of this module you can pause a graph for human review, let a person edit the state, and resume - the pattern that makes an agent safe to deploy.
Streaming and Long-Running Workflow UX
By the end of this module you can stream a long-running workflow so a user sees progress, rather than watching a spinner and assuming it has hung.
Multi-Agent Patterns with Supervisor or Handoff
By the end of this module you can implement supervisor and handoff patterns, and judge honestly whether more than one agent is earning its cost.
Debugging, Tracing, and Failure Recovery
By the end of this module you can trace a graph run, find the node where the state first went wrong, and recover the workflow rather than restarting it.
Capstone: Approval-Aware Research Workflow
In this capstone you will build an approval-aware research workflow with conditional routing, checkpointing, a human review step, and a trace that shows every decision it made.
First-release boundary
This is a text-first course shell. Leonardo/image creation, image QC, dependency-heavy runtime QA, external review, and learner pilot evidence remain separate later quality steps.