混元3D场景生成与交互大模型研究员（北京/上海）
Core
Researching next-generation 3D scene generation and reconstruction large models, and developing world models for real-time, controllable, physics-compliant interactions.
Role type
Senior IC research scientist (3D generation & world models)
Builds
3D scene generation models, reconstruction models, and interactive world models
Domain
Artificial Intelligence / 3D Graphics / Game & Simulation
Deliverable
production ML models
Required skills
Generative models (DiT, self-forcing), 3D generation, image generation, controllable video generation, 3D reconstruction (Gaussian Splatting, 3R, VGGT, MVS, SLAM), post-training (RL, DMD), model distillation, rapid prototyping, strong mathematical foundation
Preferred skills
Top-tier AIGC publications, world model expertise
Technologies
DiT, Gaussian Splatting, 3R, VGGT, MVS, SLAM, RL, DMD
Responsibilities
Develop 3D scene generation and reconstruction large models to improve quality, speed, and scale; Apply models to game and simulation domains; Combine 3D/2D generation to build interactive world models; Implement ideas with strong engineering capabilities
Seniority
Senior, hands-on IC
