Your goal: explain how LLM systems get attacked, pick the defense that actually holds, and
say how you would keep private data where it belongs.
How this works (about 22 minutes)
- Ada's 3D lab sits beside the call and shows each move as you reach it.
- Watch me: Ada works one question; you predict each move by tapping or saying it.
- Your turn: a new question. You pick where the control goes and the board follows you.
- Timed drill: a real interview question, about three minutes, with feedback and one retry.
- Your rules: your own dos and don'ts on the notepad.
Tips
- Read the module reading first: Securing an LLM System.
- Think like the owner of the system: what is reachable, and what is the worst case?