A practical 1.5-day workshop where faculty build, test, and critique AI agents. The room moves from prompting to tool use, agent loops, multi-agent workflows, and a real attendance decision system.
The published schedule is intentionally hands-on. Track B can be dropped if lab setup is incomplete; every essential concept remains on Track A.
Facilitator note: The full material is larger than the available second day. Drop Track B first, then the web-fetch lab. Keep Lab 6A, Lab 7A, the tools section, and the attendance system under pressure.
Build an Actor that reads the Excel template, identifies students below the attendance threshold, and prepares respectful parent emails.
Open the lab hub for individual exercise pages, objectives, tracks, and links to the complete step-by-step manual.
Browse every exercise from setup to capstone.
01–02Build a quality prompt and a reusable QP Generator.
03A–05AMove from plain code to Gemini feedback and structured output.
06ABuild a Teaching Assistant by hand and see the loop clearly.
07A–08AMeasure attainment, then let graders, moderators, and writers collaborate.
10ADuplicate the loop and change the tools and instructions.
11Shape an agent worth building next.
SOURCEOpen the full instructions and code.
Each book now has an HTML reading page with chapter navigation and the original PDF embedded for complete reading.
What agents are, how models work, failure modes, and prompting as engineering.
Actors, structured output, agent architecture, and tools.
Multi-agent systems, grounding, guardrails, evaluation, and practice.
Read online where it helps, and keep the original files for printing, sharing, and offline lab work.