Stage 3 · Building with AI
How AI Agents Work
Tool calling, planning, memory, MCP/A2A, sandboxes and evaluation for agents that act.
6 lessons2 ready
- 01 How AI Agents Work: The Agent Loop Model + tools + loop: observe, think, act, and the patterns that make agents reliable. Intermediate 25 min ✓
- 02 Tool Calling and MCP (Model Context Protocol) How function calling works and how MCP standardises tools for every AI app. Intermediate 25 min ✓
- 03 Agent Planning and Memory Task decomposition, reflection, short- and long-term memory and context engineering. Advanced 20 min ✎ In progress
- 04 A2A and Multi-agent Systems Agent cards, task delegation and when several agents beat one. Advanced 15 min ✎ In progress
- 05 Sandboxes, Browsers and Computer Use Letting agents run code and drive GUIs safely: containers, VMs, permissions. Advanced 20 min ✎ In progress
- 06 Evaluating Agents: SWE-bench, OSWorld and Your Own Tasks How agent benchmarks work, what their failure analyses reveal and how to test your agent. Advanced 20 min ✎ In progress
Matching units in the new course
Each unit has goals, algorithms, must-read papers, a project and a self-test.
U82 Agents: Tool Use, MCP, Planning & Memory ReAct, function calling, MCP, reflection, memory, multi-agent systems, computer use and sandboxes. 12 h10 algorithms4 paperslesson ready U83 AI Engineering: Evals, Cost & Shipping Your own eval sets, LLM judges, observability, model gateways, cost control and product design. 8 h3 algorithms1 papers