Things I've learned applying AI and data science to real problems — written when the lesson is still fresh, not after it's been smoothed into a conclusion.
The hard part of building a bank wasn't the dashboard. It was the moment three vendors all had a slightly different opinion about who the user was — and a transfer was already in flight.
I came into AI assuming the ethical questions were a slide at the end of the deck. They turned out to be the whole deck. Here's what changed my mind.
Moving 2.4M rows from MySQL to Postgres without taking the app down. The schema mapping, the gotchas, the rollback plan.
A short tour through pricing pockets, neighborhood premiums, and the seasonality of every East-Nashville two-bedroom.
When the legal answer and the right answer diverge — a small set of decision frameworks I use to keep myself honest.
Why a result being significant doesn't make it large, important, or actionable — and what to report instead.