Learning track / Intermediate / Estimated 4–6 hours
AI Agent Engineering
Explore tool selection, observations, state, and bounded execution.
Start with the engineering.
- Understand
- Inspect
- Build
- Verify
01 / Core concept
Separate the model, instructions, permitted tools, and execution loop. The model proposes; the application validates.
02 / Engineering practice
Inspect each tool observation before deciding the next action. Keep state explicit and cap turns.
03 / Test the boundary
Test unsupported tools, invalid arguments, and missing evidence; escalate instead of inventing a result.
This starter track connects published foundational lessons and guided browser workshops. Extended Python, API, retrieval, and deployment curriculum retains its existing publication status.
- Tool calling
- Planning
- Memory & state
- Execution limits
- Guardrails
- Evaluation & debugging