Build the machinery around an agent loop.
Implement tool use, planning, reflection, memory, multi-agent coordination, evaluation, and guardrails around a deterministic model stand-in.
What you will leave with
You can separate model reasoning from the orchestration, permissions, state, and tests that make an agent system dependable.
What you will be able to do
The curriculum is organized around these 4 practical outcomes.
Implement and trace an agent loop
Design typed tools and validate calls
Add planning, reflection, and memory deliberately
Evaluate multi-step behavior and enforce guardrails
Course map
Curriculum
Open a module to inspect every section before you start. Your progress follows you through the course.
Chatbot vs agent, The agent loop, Model + Tools + Instructions, and more.
Function calling, Structured outputs, ReAct, and more.
Decomposition, Plan-and-Solve, Step dependencies, and more.
Self-Refine, Reflexion, CRITIC, and more.
Statelessness, Context window, Virtual memory (MemGPT), and more.
Handoffs, Supervisor / orchestrator-workers, Shared state, and more.
Outcome vs trajectory eval, Task success rate, pass^k reliability, and more.
Caps & sandboxing, Human-in-the-loop, Prompt injection, and more.
Who this course is for
Start with the background you have. The prerequisite notes above tell you exactly what is assumed.
Engineers building tool-using AI systems
AI product teams reviewing agent architecture
Developers moving beyond chat interfaces
Continue learning
Each course stands on its own. Pick the one closest to the next task you need to handle.
Replace “it feels better” with evidence.
Engineer what the model sees and returns.
Turn one LLM call into a dependable service.
Start the course
Module 1 introduces the language and example used throughout the rest of the course.
