About
What I lead, what I build, and how I test it
I lead AI/ML products, build production agents, and publish rigorous evaluations of how they behave.
I’m a Group Product Manager for AI/ML at Capital One, where I lead product direction for ML-powered spend controls and fraud detection. I’m also the founder of Honeydew, a production family-coordination agent I began building in 2025.
Before that, I co-founded GiveTide, ran it for five years, and sold it in 2022. That arc—from founder to enterprise product leadership and back to hands-on building—is why I care about both the strategic decision and the behavior of the system underneath it.
I’m based in New York City.
The product questions
Decision boundaries
What should the system do, when should it ask, and what should it never guess?
Evidence over demos
What happens on repeated, messy, or irreversible cases—not just the best prompt in the deck?
Operating trust
How do product policy, interface design, and model behavior reinforce one another in production?
Useful economics
Where does cheaper intelligence change the product, the team, or the business model—and where does it not?
How I keep my skills current
My loop is Build → Instrument → Evaluate → Publish → Revise. I start with a real operating problem, instrument the behavior, turn failures into evaluation cases, and revise the product before I write the conclusion.
AI product leadership
Strategy, roadmaps, decision boundaries, governance, and cross-functional execution for products that depend on uncertain model behavior.
Agent evaluation & safety
Act/ask/confirm policy, destructive-action guard cases, messy-input robustness, and production-shaped evaluation design.
Multimodal systems
Voice, text, images, tool orchestration, and the product decisions that make an agent useful outside a demo.
Evidence-backed communication
Methods, limitations, negative results, corrections, and visual explanation that let readers inspect the reasoning.
AI assists with research, red-teaming, implementation, and editing. I own the thesis, source selection, evaluation design, factual claims, and final editorial judgment.
The standard is visible in the work: methods and limitations sit beside results; private user data stays out of public research; corrections remain visible; and conclusions narrow when the evidence is thinner than the first draft suggested.
Background
I earned a B.S. in Finance & Economics from 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.
ChatGPeTe
Research editions and essays
I lead AI/ML products, build production agents, and publish rigorous evaluations of how they behave—plus essays when I have something worth saying.
Subscribe to ChatGPeTe →