机器人基座模型研发科学家_CR
Core
Researching next-generation robot foundation models, world models, and world action models to advance general embodied intelligence.
Role type
Senior Research Scientist (Robotics Foundation Models & World Models)
Builds
Scalable robot foundation models, world action models, and large-scale video-action datasets for physical AI applications.
Domain
Robotics, Physical AI, Embodied AI, Computer Vision, 3D Geometry
Deliverable
production ML models
Required skills
Foundation Model research, World Model research, Vision-Geometry-Action modeling, Multimodal learning, 3D/Geometric learning, Video modeling, Robot Learning, Transformer/ViT, Latent World Models, Generative models (Diffusion/Flow Matching), Imitation Learning, Reinforcement Learning, Policy Learning, Sim-to-Real transfer, Data governance/scaling, Python/C++, PyTorch
Preferred skills
World Action Model experience, VLA research, Isaac Sim/MuJoCo/NVIDIA Omniverse experience, Top-tier conference publications (NeurIPS/ICML/CVPR/ICRA/IROS)
Technologies
Transformer, ViT, PyTorch, Python, C++, Isaac Sim, MuJoCo, NVIDIA Omniverse
Responsibilities
Develop novel model paradigms for joint learning of visual representation, 3D geometry, temporal dynamics, and physical world representation; Build and scale robot video-action datasets and training pipelines; Validate models on simulation and real-world platforms (manipulators, mobile, humanoid, industrial robots); Collaborate with global research teams and business units on technology transfer and prototype development.
Seniority
Senior, hands-on IC
