World Model Research Scientist Physical Ai

## About the role Kodiak is building AI that doesn't just perceive the world, it learns how the physics of the world works. We are developing large-scale generative world models that learn to predict realistic, physically consistent futures from real-world sensor data. This capability serves as the foundation for scalable closed-loop training, validation, and long-tail scenario generation, and is distilled into the onboard models that drive our autonomous trucks. We are looking for a research scientist to lead the design and development of world models capable of generating multi-sensor, multi-view, temporally coherent driving scenarios conditioned on actions, 3D scene context, and text. ## Responsibilities In this role, you will: - Design and train generative world models that synthesize realistic multi-camera video and LiDAR conditioned on ego trajectories, 3D scene context, and text - Research and implement conditional diffusion architectures for driving, including spatiotemporal attention, latent space design, and action-conditioned generation - Develop techniques for multi-view geometric consistency in generated outputs, drawing on neural rendering, cross-view attention, and 3D-aware generative approaches - Build methods for joint multimodal generation that maintain cross-sensor consistency between camera, LiDAR, and radar outputs - Design evaluation frameworks that measure world model quality beyond pixel-level metrics, including scenario fidelity and autoregressive stability - Scale training pipelines to learn from thousands of hours of real-world driving data across multiple sensor modalities ## Requirements - PhD in Computer Science, AI, Robotics, or a related field, with a focus on generative modeling, neural rendering, or video synthesis - Strong publication record or demonstrated research contributions in diffusion models, video generation, neural radiance fields, 3D-aware generative models, or world models - Experience with neural rendering and view synthesis and an understanding of multi-view geometric consistency - Proficiency working with multimodal sensor data (camera, LiDAR, radar) and familiarity with 3D representations such as BEV grids, voxel fields, or tri-planes - Strong implementation skills in Python and PyTorch, with experience training large generative models at scale using distributed training - Passion for building AI that understands and predicts the physical world to enable safe autonomous driving ## What we offer - Competitive compensation package including equity and annual bonuses - Excellent Medical, Dental, and Vision plans through Kaiser Permanente, Cigna, and MetLife (including a medical plan with infertility benefits) - MetLife Legal Services, Identity & Fraud Protection, Hospital Indemnity Insurance, Accident Insurance, & Critical Illness Insurance - Flexible PTO, 10 paid holidays, and generous parental leave policies - Our office is centrally located in Mountain View, CA - Office perks: dog-friendly, free catered lunch, a fully stocked kitchen, and free EV charging - Long Term Disability, Short Term Disability, Life Insurance - Wellbeing Benefits - Headspace through Cigna, Calm through Kaiser, One Medical, Gympass, Spring Health through Cigna, Rula (mental health navigation) - Fidelity 401(k) - Commuter, FSA, Dependent Care FSA, HSA - Various incentive programs (referral bonuses, patent bonuses, etc.)