About

I'm an Operating Principal at Camber Partners, a growth equity firm that invests in B2B SaaS and AI software companies. We don't embed with the teams we back. We work alongside them, helping level up their people and processes, speed up growth, and steer clear of the mistakes we've watched others make, and the ones we've made ourselves. We have the scars to prove it.

I started as a 3D artist at Kaibridge, a small game-studio startup, and worked my way up from there to CTO at Vizme, another startup. Later I went the entrepreneur route and founded Motive Arts, a generative-AI video startup that ran from 2012 to 2014. It was a little too ahead of its time.

Along the way I earned a PhD in physics, with a thesis in computational astrophysics, and then spent a stint as a paleontologist, working on a grant for the Swiss government's Paléojura project. Road construction in the Jura Mountains had uncovered dinosaur tracks, and excavations ahead of the crews found more than 15,000 of them. My colleague Kent Stevens and I were asked to analyze them. We wrote software to work out how dinosaurs walked from their fossil footprints. Before that I made dinosaur locomotion visualizations for museum exhibits and National Geographic. The archive has the details.

Eventually I was building and running data science and engineering teams, which is what brought me to Camber when it was more vision than reality. I got in on the ground floor and have been helping build it ever since.

I'm practicing agentic engineering, mostly to find out what's actually possible now. I run a software factory: I hand it a single mission, and the agents break it into many small pieces and work through them. The pieces are small. The outcome isn't. One mission can land a big change in a short amount of time.

That changes what's worth taking on. My spare time is limited, and plenty of ideas used to die because they were too tedious or would take too long to build by hand. With agents doing the grinding, I get to chase them anyway. Here's what I keep coming back to:

• Custom agent harnesses
• Multiplayer agent layers
• Custom-framework agents
• Voice agents
• Agent skills that travel across harnesses
• Model behavior experiments
• Programmable computer peripheral integrations, like the Stream Deck

The technology is ahead of our ability to use it, and closing that gap is where the fun is.

Short notes, longer writing, and standalone artifacts, which are small interactive things I built because I wanted to see the idea and not just describe it. The old project pages live in the archive.

If any of this sounds like your kind of problem, say hello on LinkedIn or GitHub.