Learning path

Agent Reasoning Patterns

ReAct, plan-and-execute and reflection with self-correction, with animated films and runnable Python. Part of the Agentic AI Engineer premium track.

Start with lesson 1

1

ReAct: Reason and Act

The pattern behind most agents: think, act, observe, repeat. Run a classic text-based ReAct loop with a safe calculator, parse actions, and see how native tool calling and…
4 min read
2

Plan-and-Execute

For long tasks, plan first and then execute: a planner writes steps with dependencies, an executor runs them in order (and in parallel where possible), and the agent…
3 min read
3

Reflection and Self-Correction

Let agents check and improve their own work: the generate-critique-revise loop, why real feedback (tests, tools, sources) beats self-opinion, a runnable code agent that fixes itself from failing…
3 min read