Chloe had built the model that examined pressure variations across older pipes.

Sophie had devised a way to identify suspicious patterns while filtering out noise from ordinary traffic and temperature changes.

Together they had built a demonstration that the judges could test against unseen data.

An engineer from a public utility was asking Sophie about the model’s error rate.

Instead of giving the polished answer a sponsor might prefer, she opened her notebook and explained which neighborhoods had not yet been represented in their test data.

The engineer nodded with respect.