Description
The role is a research position focused on embodied AI, including next-generation robot base models, world models, world-action models, vision-geometry-world-action paradigms, multimodal pretraining, imitation and reinforcement learning, and robot learning. Responsibilities include building large-scale robot video-action datasets and training pipelines, validating models on simulated and real robotic platforms, and translating research into Bosch technology and products. The posting seeks candidates with strong research experience in foundation models, multimodal learning, 3D geometry, video modeling, or robot learning, along with Python/C++, PyTorch, and English communication skills.
