I work somewhere between product and engineering, and at Ditto I don't
really separate the two. I'll write the PRD, build the thing, and then a
couple of weeks later I'm the one looking at what the numbers did. That last
part is my favorite. It's where you find out whether the idea was any good,
and it's usually more surprising than I expect it to be.
I got into all of this through robotics. Two years training perception
models for an apple-picking arm taught me how much distance there can be
between a number on a slide and a machine in an orchard, since 86% accuracy
is a genuinely good result and also means the arm comes back empty more
often than you'd like. I've been interested in that gap ever since. Since
then it's been behavioral research on how people reason about cause and
effect, and consulting work that was mostly about turning a hundred user
interviews into something a team could build.
What connects all of it is that I really want to know whether something
worked. I'd rather build something small I can measure than something big I
can't, and I'd much rather find out I was wrong early than stay confident
for months.
Outside of work I'm music director for an a cappella group and I play in the
Stanford Wind Symphony, and I spent a term at Oxford studying platform
regulation and AI ethics, which I ended up loving more than I expected to.
I'm always happy to talk about matching algorithms or four-part harmony.
See what I've built
stanford '28
