Winning 2nd at IE Games: what 48 hours of competition taught me

Forty-eight hours, one problem statement, and a solo entry. IE Games taught me more about shipping under pressure than any polished retrospective could.

The strategy

I decided not to attempt everything. I selected the three features the judges would actually experience, built the demonstration flow first, and protected that scope from every attractive distraction.

What worked

A familiar stack removed setup decisions. A simple Must / Should / Maybe board kept the project honest. Thinking about the demo before the backend forced every technical decision to serve an observable outcome.

What broke

At hour 32, the AI integration began timing out under load. The cause was simple connection pooling, but exhaustion turned a small fix into hours of diagnosis.

The lesson

Sleep is part of engineering capacity. So is writing down the reason a feature exists before the pressure begins. Both protect judgment when the clock gets loud.