UX engineer at Google DeepMind. I work on how people control robots.
Teleoperation interfaces, the infrastructure behind them, and the data collection that turns operator sessions into training data. Before robotics: AR and VR at Google and in automotive research, and games before that.
Current side project: VJ Simulator, a VR rhythm game where you control the entire DJ set.
Recent writing
Charts stopped being hardcoded and started coming out of actual mp3s, the three hero lights became a twenty-light stage rig — and the big one, there's now a chart editor that runs inside the game itself.
Sliders now spawn per-note like buttons, the room got real spotlights and lasers instead of plain cubes, and — the big one — it's playable with actual VR controllers now.
The proof of concept moved out of its flat 2D lanes and into a 3D room, and the project now has the VR groundwork in place. Still on mouse and keyboard for now.
Selected projects
Google DeepMind
Control surfaces for driving real robot hardware, and the systems that record those sessions as training data — two audiences with genuinely different requirements.
Automotive research institute
VR simulation for vehicle design research, with runtime tooling to capture the metrics that make a study possible rather than just a good demo.
Sample code and demos shipped with the Tango Unity SDK, covering mesh occlusion and ground detection.
Elsewhere
The fastest way to reach me is email.