百度公有云具身大模型研发工程师（J100650）
Core
Develop and iterate general-purpose embodied large models (VLA, World Model, World Action Model) for perception, understanding, planning, action generation, and control.
Role type
Senior IC machine-learning engineer (embodied AI)
Builds
Embodied AI models and training systems
Domain
Robotics and Embodied AI
Deliverable
production ML models
Required skills
LLM, VLM, VLA, World Model, image/video generation models, multi-modal model training, pre-training, fine-tuning, hyperparameter tuning, evaluation, Transformer, Diffusion Policy, robot control, motion control/action tracking, distributed training, multi-modal joint training, RL/Offline RL, simulation, sim2real/real2sim transfer, egocentric data processing
Preferred skills
N/A
Technologies
VLA, World Model, World Action Model, Transformer, Diffusion Policy
Responsibilities
Train and iterate general-purpose embodied large models; develop algorithms for perception, understanding, planning, action generation, and control; build data and evaluation systems including simulation and real-world data; optimize large model training systems for stability, resource utilization, and throughput
Seniority
Senior, hands-on IC