Judgment beats taste
Taste can't be argued with, I think that's its appeal. Judgment makes calls you can check later, and for whatever reason "taste" feels too pretentious to me.
15 years of experience spanning brand and product at companies large and small, most recently at Meta.
Featured work
Reactive Feed
A progressive-disclosure framework that lets feed ads respond to engagement. Adopted as a multi-team experimentation program.
Generative Landing Surface
An LLM-generated post-click surface that understands an advertiser's site and assembles the most relevant landing experience.
AI Commerce
Rebuilding post-click shopping as a native surface, with AI-powered comparison, highlights, and try-on layered in.
Encoding Design Intent
Self-built tools that turn working prototypes into engineering-ready context. Design intent as the first-class artifact.
About
Ten opinions, formed over a lifetime of filled sketchbooks. Each is a judgment, so each could be wrong. That's actually the point.
The focus point moves. The job doesn't.
There's always a spot where humans and machines try to understand each other. Today it's a screen, soon it won't be. That spot is the job.
Intent is the expensive part
Pixels, prose, and code are nearly free. What's scarce is knowing what to make and saying it clearly enough to build from. I've been practicing this stuff for a while.
Simulation beats empathy
Empathy is hard to verify. I run a mental movie of a stranger's first encounter. What they notice, misread, or distrust, then build it and learn if I was right.
A prototype is the truth arriving early
A small piece of the future, delivered early for inspection. Wrong here costs an afternoon, wrong in production costs more.
Build the first version yourself
The most persuasive document is the thing itself, working, in someone's hands. It works every time.
The team's fluency is the job
Being AI-native designer is a good idea these days. Making the team AI-native is the real work. Being the only rower on the lifeboat is a poor plan.
Calm is a feature
Panic never improved a design decision, and someone in the room has to think the weather is fine. My managers and peers have always said I'm that guy.
The obvious question goes unasked
Smart rooms skip the simple question, assuming it's answered. Asking costs four seconds of embarrassment and pays for the whole project.
Short loops beat big rooms
Good work tracks how fast a decision meets its result... in days, not quarters. Small teams, tinkering, treated as work. The best work happens there.