Physical AI Engineer (Model)

About this role

Join our AD Division to develop generative AI models that control real-world robots and autonomous vehicles. You'll create end-to-end trajectory generation and decision-making systems using reinforcement learning, imitation learning, and motion planning techniques, deploying safe and efficient Physical AI solutions in complex driving environments.

What you'll do

• Preprocess data, train models, and validate performance for end-to-end trajectory generation and decision-making systems
• Develop and optimize model-based reinforcement learning and imitation learning algorithms
• Integrate and test motion planning, filtering, and navigation algorithms
• Optimize deep learning models and planning pipelines using CUDA for real-time on-device performance
• Validate AI models and algorithms in simulation and real vehicle environments
• Collaborate with perception, planning, and control teams on next-generation Physical AI systems

What they're looking for

• Python and modern deep learning frameworks (PyTorch, JAX)
• Reinforcement learning and imitation learning
• Motion planning, filtering (Kalman, Particle), and navigation algorithms
• Machine learning and deep learning
• Simulation and hardware validation
• C++, CUDA, and TensorRT optimization
• End-to-end trajectory generation
• Autonomous driving or robotics AI development