Join Us

Job Responsibilities

1. Research and develop embodied intelligence algorithms for robotics, including but not limited to: VLA , imitation learning, reinforcement learning, and multimodal perception/fusion (vision, speech, tactile).

2. Build multimodal embodied operation data collection systems (simulation + physical robots), constructing datasets and algorithm validation benchmarks.

3. Participate in model architecture design and training, leveraging SFT /RFT to enhance performance; track cutting-edge research and propose innovative solutions.

4. Optimize model inference deployment systems to achieve precise robotic manipulation in complex environments, ensuring high operational efficiency.

5. Design and optimize systems for specific scenarios (e.g., dual-arm collaboration, complex object manipulation, tactile feedback, HRI), driving VLA-based solutions to real-world robotic platforms.

6. Proactively identify and analyze client needs, expanding practical application scenarios to maximize the value of embodied intelligence algorithms.

Requirements

1. Master’s degree and above (or Bachelor's with exceptional profile) in CS, AI, Robotics, Control Science, or EE; domain knowledge in CV, NLP, or multimodal learning.

2. Proficient in Python/C++ with solid data structure/algorithm foundations; experience with PyTorch/TensorFlow/JAX; familiarity with main VLA and imitation learning models (e.g., Diffusion Policy, ACT, PI0, GROOT); RL knowledge.

3. Hands-on robotics project experience; skilled in IsaacLab/Isaac Gym/MuJoCo; knowledge of Sim2Real/Real2Sim techniques.

4. Ability to deeply comprehend cutting-edge literature, with sustained passion and tracking of industry advancements; strong collaboration/communication skills; publications at top venues (ICLR/CVPR/ICRA/CoRL) is a strong plus.

Note: This position is also recruiting interns.