About

What I build, what I test, and why I write

I work at the point where AI leaves the demo and meets someone's actual day.

I split my time between two roles: Group PM for AI/ML at Capital One, where I lead work on ML-powered fraud and spend management, and founder of Honeydew, an AI-powered family coordination app that uses voice, text, and photos to turn the day's loose ends into shared plans.

Before that, I co-founded GiveTide, ran it for five years, and sold it in 2022. Building that company the expensive way is what made the new economics of AI-assisted software impossible for me to ignore.

"Intelligence became abundant. Intention did not."

The Questions I Keep Returning To

Product Judgment

What should the system do, when should it ask, and what should it never guess?

Evidence Over Demos

Real prompts, published methods, and the bad results beside the good ones.

Tools That Coordinate

How an agent chooses among 60+ tools without becoming reckless or useless.

Human Work

What gets more valuable when prose, code, and pattern recognition get cheap?

Background

I studied Finance & Economics at the University of Richmond, graduating Magna Cum Laude. My career has moved through fintech, fraud systems, consumer software, and the practical limits of AI agents. The through-line is less glamorous than “innovation”: take an ambiguous problem, make the tradeoffs explicit, and ship something people can trust.

I write the work down.

The essays range from model benchmarks to the parts of judgment a model cannot supply.

Follow on Substack →