Beyond the resume

What keeps me here is the distance between what a model can do and what it actually changes for a person. Most of the interesting work lives in that gap — framing the problem so it matches a real decision, being honest about what the data can and can't support, and shaping the output into something a person can act on and argue with.

That's also where responsible AI stops being an abstraction for me. A system nobody can interrogate doesn't earn trust, and it shouldn't. I'd rather build something a domain expert trusts enough to push back on than something that scores better and sits unused.

The version of AI I want to work on augments the person using it: the underwriter, the teacher, the researcher still owns the decision, and the system makes it faster and better informed. Getting there is less about raw capability than about the unglamorous parts — what the system explains, where it defers, and how it fails.

Most of what I know came from people who understood a domain far better than I did: proteomics researchers who could tell a real spectrum from a plausible one, teachers who tuned problem difficulty with me grade by grade, underwriters explaining why a rule that looks arbitrary from the outside exists.

Moving across bioinformatics, education, and finance meant starting over as the least knowledgeable person in the room every couple of years. I've come to like that. It forces the basic questions, and the basic questions are usually where a project's real problem is hiding.

I enjoy that part of the work as much as the building — arguing about what a metric actually measures, or about where a model should stop and a person should take over.

I travel whenever I can — twelve trips and counting, logged in a running Notion database with country, dates, and region, because apparently even my vacations get a schema. The parts I remember are rarely the ones on the itinerary, though — usually a long meal with people I hadn't expected to be sitting with. Sports and music are how I reset.

I like listening to people's stories. A fair amount of what I find worth building comes out of hearing how someone actually spends their day, and it's the fastest way I know to find out that a problem I assumed was solved isn't.

Mostly I try to stay attentive to small things: a good conversation, an unexpected kindness, an ordinary afternoon that turns out well. That's where most of the energy I bring to work comes from.