The flight plan wasn't linear.

I got my pilot's license before my first apartment.

Private, commercial, instrument-rated. The whole progression, done young and done fast.

There's something about flying that rewires how you think. You learn to read instruments under pressure, navigate complex systems in real time, and make decisions when the margin for error is literally zero.

"Turns out the job description for flying and the job description for building AI are the same: navigate complex systems, read instruments under pressure, make decisions under pressure with limited information."

What flying actually teaches you.

Cross-country at 12,000 feet. Stall training. Night flying where instruments matter more than intuition.

That's where System 1 vs System 2 became practical, not theoretical. Recognize when instinct is useful, then deliberately switch to checklist-driven reasoning when the stakes are high.

Different domain, same operating system I use in AI today. I didn't know it at the time, but becoming a pilot and training in the Air Force was the best start to my career I can ask for. I didn't get to start at the start line. I started hundreds of miles behind and in the mountains, and on the way, I saw and learned so much more.

Business as a systems problem.

Krannert taught me to break business problems into testable pieces and make decisions with data, not vibes.

Case competitions were training: solve fast, defend your answer. Won a Boeing case, and got second place in the next year's Boeing case. Also got second place in Indianapolis University case. I won and lost for the same reason — I was creative. I was not afraid to reinvent the business and start new ventures, while others slapped on the classical business-school frameworks. Those moments made it clear: Analyzing went only as far as the system allows you to go. To truly unlock myself, I needed to build opportunities.

From auditor to data engineer.

Started by auditing financial statements at Crowe. Rigorous, high-accountability, and repetitive enough that I kept asking how much of it could be systematized.

That question pushed me toward automation and into EY as a data engineer, where I built pipelines that reduced manual work. This was before "data science" was a thing, but the pattern was already there: instrument the process, automate where you can measure quality.

ML at global scale, under pressure.

I joined Uber at one of the most intense moments in the company's history. The London license was threatened. Regulatory pressure from every direction. The company fighting for survival on multiple fronts. I worked with some of the most driven colleagues in saving our UK business.

Watching the machine learning systems at Uber work at scale was a phenomenal sight. Not in theory, not in a notebook. In production, with millions of predictions that need to be fast and accurate. The decisions we made at Uber didn't only affect "drivers" and "riders". Our decisions would affect the lives of dads trying to provide for their family because driving was their job, and moms trying to make an extra buck to buy soccer shoes for their son. That's where I learned what it really means to ship models that matter. Not to satisfy numbers, but to affect humans.

"The gap between ML that works on your laptop and ML that works at Uber scale is enormous. Closing it changed how I think about everything."

Master's in data science, while working full-time.

Deep learning and scalable data systems, earned in evenings and weekends while shipping Messenger features at Meta by day.

One capstone: a model that rewrites toxic content without losing the original meaning. Hard problem. If there's ever a good place to catastrophically fail, I advise failing in school. I tried to invent a novel U-net architecture, training the model to diverge in learning semantic and style with a new loss function. What seems intuitive and basic today was a learning experience in 2021.

What I still carry from MIDS: connect research to product outcomes, make quality measurable, design systems teams can actually ship and trust.

Shipping ML to billions.

At Messenger I shipped high-quality zero-to-one launches: technical depth meeting product judgment, working closely with engineers and PMs.

By 2024 I was leading MetaAI's largest Messenger update: from Tab to Search integration and beyond.

The frontier.

At Meta Superintelligence Lab, I moved to the bleeding edge: foundation-model quality, data curation, evaluation systems. Same cabin discipline, at model scale.

"Most AI writing is by people who've never trained a model. Decision Height is by someone who has."