World Model Research Scientist- Physical Ai

Research Scientist in Physical AI at Kodiak Robotics, focused on designing and developing large-scale generative world models. This role involves creating AI that learns physics to predict physically consistent futures from sensor data, enabling scalable training and validation for autonomous trucks.

Skills and focus areas

• generative world models
• multi-sensor data
• multi-view video
• temporal coherence
• conditional diffusion architectures
• spatiotemporal attention
• latent space design
• action-conditioned generation
• real-world sensor data
• autonomous driving