1、Research high-fidelity dynamic world models, overcoming challenges in environment prediction, multi-modal dynamics modeling, and uncertainty quantification. 2、Build decision systems combining world models and reinforcement learning for efficient planning and control. 3、Optimize the model's generalization and computational efficiency, addressing drift issues in complex environments, and pushing for the practical application of robotics and autonomous driving. 4、Track cutting-edge research from ICLR, NeurIPS, CoRL, leading architectural innovation and theoretical analysis.

1、PhD or excellent Master’s in computer science, AI, robotics, or related fields. 2、Expertise in dynamics modeling, temporal prediction; familiar with architectures like Transformer, diffusion models (e.g., Dreamer, World Models). 3、Experience combining world models with RL, able to independently verify the effectiveness of prediction and decision-making. 4、 Experience with top conference papers or open-source projects preferred. You Will Gain: 1、Contribute to the construction of industry-leading world model systems, exploring the core path of general intelligence. 2、Access interdisciplinary resources (e.g., cognitive science, control theory), supporting both theoretical and engineering growth. 3、 Implement closed-loop validation and iteration in robotics and simulation platforms, pushing for practical application.