Graph engineering academy

From agent loops to intelligent workflows

Seven modules, about 66 minutes, no graph theory assumed. You start by watching one agent decide which tool to call, end by choosing the least complex architecture for your own problem, and build and break a graph on the way. One idea runs through all seven: a graph is not more powerful than a loop. It writes down structure the loop leaves to the model, which helps when a job needs that structure and costs engineering when it does not.

  1. 01 What is an agent loop? What happens when a model decides which tool to use, again and again? Name the four beats of a loop: observe, decide, act, receive feedback. Tell a model decision from a tool running. Say what stops a loop and why the model should not be the only thing that can. 6 min
  2. 02 What is a graph? What are nodes, edges and state, without any graph theory? Read a graph: nodes do work, edges say what runs next, state is what they share. Build a three-node workflow and run it. Spot a cycle and the condition that ends it. 8 min
  3. 03 From a loop to a graph What changes when the same job is a graph instead of a loop? Compare two architectures on one task. List what a graph makes explicit: control flow, handoffs, state, checkpoints, approval, recovery. Explain why a graph is not more powerful than a loop, only more explicit, and what that costs. 10 min
  4. 04 The graph playground Can you build a change-request team and break it? Add nodes, connect them, define branches. Add a tool, an agent, a reviewer and a checkpoint. Introduce a failure and replay the run. 15 min
  5. 05 Graphs in the real world What do graphs look like in software development, research, support and security? Recognise seven common graph shapes. Say where each one waits for a person. Pick the shape closest to your own workflow. 10 min
  6. 06 Production graph engineering What does it take to run a graph for real? Name the eleven properties a production graph needs. Explain idempotency and why resume depends on it. Plan how a graph is versioned while runs are in flight. 12 min
  7. 07 Choose the right architecture What is the least complex thing that will work? Walk the decision from "is an agent needed?" to "is multi-agent justified?". Defend the simplest architecture that meets the requirement. 5 min

By the end you should be able to explain the difference between an agent, a loop, a workflow and a graph to a colleague, and say which one their problem needs.