I’m building an online fashion experience where a shopper snaps a couple of photos, instantly receives a full-body 3D avatar, and can try on any garment or accessory in our catalogue from either a phone or a desktop browser. Here’s what I need put in place: • AI-driven pipeline that converts 2–3 user photos into an accurate, skinnable full-body avatar (no markers or depth cameras). • Web-ready rendering layer—WebGL, Three.js, or similar—that lets the same avatar load smoothly on mobile phones and desktop computers. • Garment draping and collision logic for tops & shirts, pants & skirts, plus hats and accessories, with realistic cloth behaviour and texture fidelity. • Simple uploader or API so my team can drop new clothing meshes (OBJ/FBX/GLB) and see them auto-rigged to the avatar skeleton. • Output of fit images or a short turntable video that users can download or share. Acceptance criteria 1. Avatar generation completes in under 60 seconds on a mid-range phone and standard laptop. 2. Size and body-shape accuracy within 1 cm tolerance from supplied measurements. 3. Garment try-on runs at 30 fps or above on recent iOS, Android, and Chrome/Edge desktop browsers. 4. Code is modular, well-documented, and delivered with a short demo project I can host on my own server. If you’ve worked with Blender scripting, Unity or Unreal cloth, TensorFlow, PyTorch, or similar toolchains and can demonstrate a working sample, I’m ready to get started right away.