I’ve spent my career at the intersection of the engineers who can build anything and the end users who know exactly what they need — translating between the two and turning that into products people rely on. AI tooling turned that vantage point into a build capability of its own.
I don’t guess at requirements. I sit with the people doing the job, watch where the workflow breaks, and write the spec in their language before anyone touches a keyboard. Every product on the previous page started as a problem a real user described to me.
Vibe coding gets a feature on screen. Shipping it is a different discipline. I direct the agents the way I’d direct an engineering team: clear acceptance criteria, review of every change, performance and security passes before anything goes live — and I make sure I understand the ins and outs of what was built, not just that it works.
If a product can’t tell you whether it’s working, it isn’t finished. I define the metric first, then build toward it — which is how I ended up creating an industry-standard score for email deliverability instead of just reporting on it.
Every team has two groups that struggle to talk to each other: the people who can build, and the people who know what they want. I’ve always been the one who could hear a sales rep describe a frustration and hand engineering a spec they could build from — and hold the finished feature to that standard.
Large language models and agentic tools let that translator also do the building. That’s the story of the portfolio: the same product instincts, with the delivery bottleneck removed.
What that looks like day to day: I write the brief, I break it into features an agent can execute, I review the output as a product owner and as the person who has to support it at 2 a.m., and I ship.
Created the idea, the formulas and the algorithms behind SUTOR, a trademarked industry metric for email marketing and deliverability scoring.
Directed and product-managed the architecture and creation of a data warehouse from the ground up — defining what needed to be measured, how it should be modeled, and what the business would do with it.
Specified, built and shipped with AI agents, spanning DNS and email authentication, inbox placement testing, developer utilities and a handful of things built purely for fun.