I Shipped It. Then I Learned How It Worked.
Building a real product with AI, as a complete backend beginner

A women-in-tech nonprofit needed a platform to sell event tickets. I said yes, a frontend engineer who had never built a backend, never handled payments, never deployed to the cloud. With AI writing most of the code, it shipped: real users, real money. Then things started breaking.

In this workshop:

• Three real bugs from a live system: ticket counts that reset themselves, a free event with no capacity limit, confirmation emails that fail silently. One is still in production today.

• The spec I didn't want to write: I wanted to start building immediately. A senior engineer insisted on a spec first. Every bug above traces back to a question that document didn't ask.

• How to interrogate an AI's answers: what to ask before you accept generated code, and why a plausible answer still isn't enough.

We'll walk through the real product, the real spec docs, and the code behind each bug. You'll leave with a one-page checklist for your next AI-assisted project.