I've always needed to understand how things work, and once I do, I want to
build something with it. That's what took me to cognitive science: it
studies the most interesting problem-solving system I know of,
the human brain. I spent two years at UCLA studying it up
close, running behavioral research in the lab. Software is how I turn that
understanding into things people actually use. The four apps above came out
of that, shipped end to end, each one an experiment in the cognitive
psychology of its users.
LLMs came into my life as a personal tool. Before I ever shipped a product
with them, they were helping me learn and write, even cook and practice
music. They opened my eyes to how much of a life this technology could
reach, and I've been building with it and around it ever since: agentic
coding tools for development, generative pipelines for
creative and research, automation wherever it adds leverage. I want to
understand how people work, and I build products that put that
understanding to the test.
Read the full version
I love building things and solving problems. Cognitive science pulled me in
because it studies the most interesting problem-solving system I know of —
the human brain, and how it works underneath to produce the whole of
human experience — including the way we live with the tools and
technologies our culture keeps producing.
Two years of lab work at UCLA's TLC Lab, alongside coursework in memory,
neuroscience, decision-making, attention, and consciousness, shaped the way I
work more than anything else: operationalizing experience —
turning something as soft as attention or motivation into parameters you can
track, test, and be proven wrong by. Consumer products are the same problem with worse
instrumentation. A paywall is a hypothesis about willingness to pay; an
onboarding flow is a hypothesis about attention and friction. I ship them as
tests and read the results honestly.
The same lens runs through the funnel above the product. An ad creative is an
attention problem before it's a marketing one — what stops a scroll, what earns
the next second, what carries someone from a paused thumb through install,
onboarding, and paywall without breaking the thread. I design for that whole
path: grabbing attention, protecting it, and treating the ad
and the product as one continuous experience.
In college I started using ChatGPT to automate and speed up my own work, and
fell in love with the technology. What fascinates me is what it means, as a
human, to have intelligence available at hand at all times — and how much
skill there is in using it well. Working with these models is a discipline of
its own: directing that intelligence, communicating what you need efficiently
and consistently, and evaluating what comes back systematically rather than
taking it on faith. The model is the ultimate extension of
yourself, and the quality of that extension depends on how precisely
you can articulate what you want.
The leverage compounds through learning. New capabilities arrive almost daily,
and staying on top of them — consistently, as a practice — is part of the job.
So is the other half, which moves partly on its own: deliberately looking for
new places and ways to put AI to work in your life, rather than waiting for
the models to improve. The best work of the coming years won't necessarily be
made by AI, but it will certainly be made by humans using AI:
focusing on what actually matters and handing the rest to the machine. That's
how my four-person studio operates — agentic coding tools for development,
generative pipelines for creative and research, and automation that lets us
work like a much larger team. That, to me, is the real promise of this
technology: LLMs are the ultimate enabler of individual
potential, opening horizons of capability that simply didn't exist
for one person before.
What keeps me hooked is that cognitive psychology and neuroscience offer a
model for nearly everything in a life — why music moves us, where creativity in
art comes from, how we think about what we eat. Evolved psychological and
neurological mechanisms sit underneath the entire system. Products are simply
the newest place those mechanisms show up — and one of the first places you can
measure them at scale.
