Senior AI & Computer Vision Engineer - Interactive 3D Visualization

We are looking for a senior developer who combines hands-on AI and computer vision experience with real 3D visualization skills. This is a practical development role. We need someone who builds and delivers working things, not someone who only talks about approaches. There is potential for ongoing work beyond the initial scope for the right person. What you will be working on Building an AI-powered pipeline that processes real-world photographs and produces a textured interactive 3D model as output. The pipeline will be embedded into an existing Windows desktop or MAC software platform as a visualization and communication feature. Required skills - Hands-on experience with SAM or similar AI segmentation models on real-world imperfect images - Image processing in Python - OpenCV, PyTorch, or similar - Multi-image texture projection and blending - UV mapping and texture baking - Template mesh deformation and fitting - Three.js - interactive viewers with per-object selection, highlighting, and annotation overlays - GLB and glTF workflow - ONNX Runtime for local inference on Windows - Comfortable in both Python and JavaScript - Experience delivering working end-to-end prototypes Experience we are looking for - You have built a real computer vision pipeline on messy real-world images - glare, occlusion, inconsistent lighting - not just clean datasets - You have shipped an interactive 3D viewer with per-object interaction, not just a static scene - You have delivered complete working prototypes, not just isolated components - Experience with imaging or visualization software platforms is a plus Technical requirements - Full pipeline runs locally on a standard Windows PC - No cloud dependency or dedicated GPU required for core functionality - GLB as the preferred 3D format - Electron for Windows desktop packaging What we want to see in your proposal Answer these two questions at the top: 1. Describe a real project where you used SAM or similar on difficult real-world images. What were the challenges and how did you handle them? 2. Describe or show a Three.js or WebGL project with per-object interaction and overlays. Concrete specific answers only. Vague claims without real examples will not be considered. If the initial engagement goes well this becomes an ongoing collaboration on a growing visualization product.