ARG Studio is where I make apps, explore new ideas, and work with people on things worth building.
The work moves between design, software development, and AI research. Sometimes that means a small, opinionated app. Sometimes it means a custom system for a client. Sometimes it starts with a question and a prototype to see where it goes.
A few things in orbit
FOOM is a family of fitness apps built around showing up. Saylight brings private, offline dictation to the Mac. Harness explores small interfaces for working with AI agents. Algorithmic Research Group develops benchmarks, datasets, and open infrastructure for AI safety research.
The creative archive reaches into generative design, data sculpture, and live performance: algorithms that grow forms, emotion data carved into oak, and an opera that runs on its own.
Different subjects, shared instincts: pay attention to the details, make something useful, and leave room for discovery.
The person behind it
I teach AI at Duke, with courses in applied AI, AI products, and software development. Previously, I was a Senior ML Engineer at Apple and worked at several startups, building software that had to hold up in everyday use.
My research focuses on AI systems and agents. I've raised over $1.5 million in research grants and presented at NeurIPS and ICML. Research informs how I build; building gives me better questions to research.
Small on purpose
I stay close to the work, from the first sketch to the code that ships. When a project needs more depth, I bring in collaborators with the right expertise.
For clients, that means a direct working relationship, visible progress, and honest conversations about what to build next. For the studio, it means enough space to pursue our own ideas, too.