Contributor

Andrei Vince is a Computer Engineering undergraduate at UNC Charlotte (Dec 2027). He co-developed a structure-aware pipeline that couples 3D cellular-automata voxel massing (64×64×64) with SDXL + ControlNet (Canny) to render photorealistic images while preserving geometry. The system automates a CA→rendering→LLM-metadata loop, producing 15,000+ unique configurations and averaging ~70 seconds per model on a single NVIDIA A40. Interests include procedural generation, computational design, and scalable dataset tooling. Presenting at SIGGRAPH Asia 2025 Posters.