Personal details
Clinton C.
Technical skills
About
I build software that runs on real hardware. Most recently that's been AR headsets for defense and industrial use: headset calibration, camera access, hand and object tracking, on-device ML and voice commands, built in Unity, StereoKit, Android native and C++.
Before that I built HoloLens 2 inspection tools and VR alignment and tracking safety systems. Before that I spent 8 years at Boeing in robotics programming and manufacturing R&D: OpenCV inspection work, a patent, and a lot of respect for things that have to work on a factory floor.
Through Arc I've been rebuilding a legacy scientific imaging app for a coral research lab: X-ray densitometry, 3D scan registration and a prototype ML model to guide band tracing. It's now in use in the lab. I also build my own projects in voice, vision, embedded and web, and I reuse that work on client projects.
I build with AI coding agents, using a planning, testing and audit process I put together myself. It gets me up to speed on unfamiliar SDKs fast. I do the testing and debugging on the device.
Available from November 2026 for perception, computer vision, XR or edge ML work, full-time or contract. US citizen living in Spain (CET), with flexible overlap into US Pacific hours.
Work Experience
Contract engineer on the perception and applied software teams for AR headsets used in defense and industrial settings, building features, capabilities and the demos that show them off.
Areas: headset and camera calibration, camera access and perception tooling, hand and object tracking, on-device ML, voice command interfaces, and device profiling. Unity, StereoKit, OpenXR, Android NDK, C++, TensorFlow Lite.
On-site in San Jose from April to July 2026, then remote from Spain. Contract through October 2026.
Part-time contract through Arc. Rebuilt a legacy VB6 and Fortran research application as a .NET desktop app, then took it well past the original. It measures growth, density and calcification from X-ray images of coral skeletons for climate research, and it's in use in the lead scientist's lab ahead of a public release.
• Image analysis: density calibration from wedge standards with drift warnings, and automated band detection with four boundary methods
• 3D scan registration: reads PLY, STL and OBJ scans of coral slabs, registers them to the X-ray, and corrects density for the real thickness
• Prototyped a small U-Net (ONNX, CPU inference) that predicts growth direction across the X-ray to guide band tracing
• Experimental scatter correction, a batch and command-line pipeline, and reports with processing provenance