Week 18: Build and Defend an AI System

The capstone is not a model demo. It is a system defense.

You will assemble the evidence developed across the applied chapters: data assumptions, model behavior, evaluation, architecture, deployment constraints, failure paths, safeguards, and human authority.

Then you will test the system under normal and failure conditions in the AI System Defense Lab.

Your final defense should make three things clear:

1. what the system is designed to do;
2. what evidence supports the design;
3. where the evidence stops.

A strong capstone is not the one with the biggest claim. It is the one where every important claim can be traced to a test or clearly marked as a limitation.