iOS Vibecoder — 3D Face Analysis App

Type: Contract or co-founder track · Stack: Swift, ARKit, Metal, Core ML What we're building An iOS app that scans the face in 3D, measures golden-ratio proportions to sub-millimeter precision, and renders a volume heatmap showing where one facial half sits lower than the other. Real geometry, real depth data — no gimmick filters. What you'll do Build the capture pipeline on TrueDepth (multi-frame fusion, scale reference, ICP registration). Turn point clouds into a clean, watertight mesh and paint a live volume-difference heatmap (Metal/RealityKit). Ship landmark detection + golden-ratio math that's honest about its accuracy. Move fast: prototype, measure, iterate. You vibecode You build with AI, not around it. You prompt, generate, test, and refine in tight loops. You'd rather have a working prototype by Friday than a perfect spec by next month. Cursor / Claude Code / Copilot are part of your hands, and you know when to trust the model and when to read every line yourself. You should know Swift + iOS, hands-on with ARKit/TrueDepth or 3D/depth work. Comfort with 3D math: meshes, ICP, signed distance, symmetry planes. Bonus: Metal shaders, photogrammetry, Core ML/Vision. How to apply Send a link to something you built fast — repo, demo, or a 60-second screen recording. Tell us one thing you vibecoded that surprised you.