I followed where the tools were going.

I spent ten years in architectural practice: federal and mission-critical projects, international work. In 2019, before LLMs, before Midjourney, before any of this was obvious, I went back to school for computer science. I didn't leave architecture. I followed where its tools were going.

That path ran through graduate CS coursework at Georgia Tech, a Masters of Computation Design for Architecture at IAAC, research published at ACADIA and CAADRIA, and work exhibited at the Venice Biennale in 2025. Since 2022 I've kept a weekly public record of machines learning to imagine architecture at @diffusion_architecture.

Today I'm Head of R&D at xFigura, building AI tools that architecture firms use every week. My research looks at what comes after prompting. Today's tools answer a prompt with an image and forget everything in between. The next ones will work the way design actually works: propose options, test them, keep what was learned, and build on it, in a loop. What the loop cannot supply is the judgment of what's good. That stays human, and everything else rests on it.

This site is the sketchbook: generated images, videos, interactive 3D, and essays on where this is all going.