具身智能数据算法专家-火山引擎
Core
Building high-quality datasets for VLA/VLN models to train embodied AI robots, covering real-world data collection, simulation, and data generation.
Role type
Senior IC embodied AI data algorithm expert
Builds
Datasets for VLA/VLN models, simulation environments, and data pipelines for humanoid robots
Domain
Robotics / Embodied AI / Simulation
Deliverable
production ML models
Required skills
VLA/VLN model data requirements, dataset construction, simulation environment setup, Python/C++, multimodal data fusion, teleoperation data collection
Preferred skills
Generative AI data generation, Sim2Real techniques, real-world robot data collection experience
Technologies
Isaac Sim, MuJoCo, action capture suits, VR equipment
Responsibilities
Design data collection standards for real robots using teleoperation systems; build simulation environments and core tasks for automated data collection; process raw data (cleaning, labeling, augmentation) to create training-ready datasets; collaborate with algorithm and hardware teams to iterate data strategies based on model feedback.
