I build immersive apps for headsets and the browser in Unity and WebXR. I also make the tools around them: computer vision that reads people and poses, and AI tooling that lets agents work inside 3D software. Earlier I built IoT sensors, browser games and gesture interfaces, and I still enjoy anything that turns a camera or a sensor into input.
Text-first Blender control for AI agents: scene diffs instead of screenshots, node trees as JSON, 104 tested recipes, and a reference-image pipeline that measures views and scores the model against them. The donut above was built that way.
Python Blender MCP MediaPipe
|
A pixel-art koi pond for GitHub profiles. The koi swims, eats, turns and grows with your public activity, and wears the colours of your languages. It ships as a GitHub Action with its own sprite sheets, and the koi further down this page is drawn with it.
Python GitHub Actions Pixel art Pillow
|
