Senior AI / CUDA / MLOps / 3D Graphics Engineer – Human Body Reconstruction Platform

Project Overview We are developing an AI-powered human body reconstruction platform that generates realistic textured 3D human meshes from user photos. Current Technology Stack: SMPL-X PyTorch CUDA OpenCV FastAPI Docker OBJ Mesh Generation UV Mapping Pipeline Texture Generation Pipeline Project Status The platform is approximately 100% functionally complete and operational. Core reconstruction, mesh generation, UV mapping, texture generation, API infrastructure, and deployment components have already been implemented and are working successfully. We are seeking a senior engineer who can quickly understand the existing codebase and focus on: Performance optimization GPU acceleration improvements Reconstruction quality enhancement Production readiness Scalability and deployment improvements The primary objective is not building the system from scratch, but significantly improving speed, quality, reliability, and maintainability before production deployment. Key Responsibilities Optimize PyTorch and CUDA inference performance Improve GPU utilization, memory efficiency, and throughput Enhance SMPL-X reconstruction accuracy and stability Improve pose estimation and pose transfer workflows Optimize mesh generation pipelines Improve UV mapping and texture generation quality Optimize FastAPI backend performance and architecture Improve Docker deployment and infrastructure automation Support cloud GPU deployment and production readiness Conduct code reviews and recommend architectural improvements Required Skills: Python PyTorch CUDA Computer Vision FastAPI Docker Linux Preferred Experience SMPL / SMPL-X Human Mesh Recovery (HMR) Human Pose Estimation AWS or GCP MLOps and GPU Infrastructure 3D Mesh Processing UV Mapping Pipelines Texture Generation Systems What We're Looking For An engineer with proven experience in AI, CUDA optimization, computer vision, and 3D graphics pipelines who can identify bottlenecks, improve reconstruction quality, optimize GPU performance, and help prepare the platform for production-scale deployment. Please include relevant project experience, GitHub repositories, portfolio links, or examples of similar work in your proposal.