Build
Write prompts and wire widgets into working AI agents in a hands-on sandbox.
Vanderbilt University · AI Program
Vanderbilt's hands-on AI course. Write prompts, build agents, test and govern them, and ship to a real mission — learning by doing, the way AI is actually built.
Learn to build AI you can trust.
How you'll learn
One end-to-end arc. You build AI in a sandbox, make it safe and reliable under real guardrails, ship it to a live mission, and see the outcomes tracked — the way AI is actually built and run.
Write prompts and wire widgets into working AI agents in a hands-on sandbox.
Make your AI reliable and safe: citation-backed claims, human-in-the-loop, no auto-submission.
Ship to a real mission domain — veteran & military benefits navigation.
See outcomes and service tracked, the way a real program reports.
How this course teaches
You don't read about AI — you use it. Adjust an input, watch the model react. Write a prompt, run it, see what comes back. Every concept is something you touch.
You assemble working AI agents through a guided interface, so the focus stays on what the AI does — not on boilerplate or setup.
You learn governance hands-on: claims need citations, a human stays in the loop, and nothing auto-submits. Safe AI is a skill you practice here.
Every build, prompt, and decision you make is recorded, so you can review your own work and understand exactly how your AI behaved.
The Mission
This isn't a throwaway sandbox demo. You build toward veteran & military benefits navigation — a live, meaningful domain serving people who need clear answers. The work you do here is the kind of work that reaches real people.
Everything you build is governed and audit-logged, so you practice shipping AI the way a responsible, accountable program actually would.
The Guardrailed Classroom
Jump into the first lesson and start learning AI the way it's actually built — by doing it yourself.