An AI platform for the back office. Built on the thesis that LLMs need frameworks to guide them, contain them, and make them operate reliably. Not another chatbot wrapper.
Fifteen years of computational biology; two companies built around the same instinct. Form follows function. The most exciting work lies where complexity emerges.
I build things at the intersection of biology, computation, and people. Lately the work has drifted from one kind of complexity to another. From predicting the structure of proteins to predicting the structure of organizations.

An AI platform for the back office. Built on the thesis that LLMs need frameworks to guide them, contain them, and make them operate reliably. Not another chatbot wrapper.
AI-driven protein engineering. madi™, a deep-learning platform, designs functional proteins for food, agriculture, and health. A decade-long arc; platform in production with Fortune-500 partners.
I keep returning to systems that resist a single explanation: biology, beauty, entropy. Something hidden ties them together. From how the first proteins found their functions, to why we feel beauty almost as second nature.
The universe is magnificent. Its scale unfathomable. That we can capture a piece of it, see it with our own eyes, and sit with the beauty of it: still incredible to me.

I'm a longtime RTS player. StarCraft, mostly: Master league, 250 APM at my peak. Ever since, I've seen automation and optimization as a real-life game. With Claude Code and the new wave of AI tools, the same instinct runs deeper than ever. Humans aren't meant for repeat. We crave creativity and complexity, and that's the engine behind every automation worth building.