AI Research Intern – Predictive World Model

What you'll do

• Lead a research project from problem formulation through architecture design, training, evaluation, and empirical analysis.
• Research future prediction, multimodal latent representations, cross-embodiment generalization, and evaluation methods for world models.
• Collaborate on scalable training and inference pipelines and publish or open-source results where appropriate.

What they ask for

• Currently pursuing a PhD in Engineering, Computer Science, or a related field, focused on Deep Learning, Computer Vision, or Generative Models.
• First-author publications at top-tier venues; work under submission may be presented as an arXiv preprint.
• Strong generative modeling and experimental methodology, with hands-on model experience in PyTorch or JAX; strong Python and software design skills.
• Available for at least 12 weeks and on-site work at the Santa Clara office.

Nice to have

• Experience with generative models for video or 3D, world models or learned simulators, or model-based reinforcement learning and VLA models.
• Experience with multimodal foundation models, video tokenizers or VAEs, autonomous driving, robotics, embodied AI, or widely used open-source projects.

Tools

PyTorch, JAX, diffusion, flow matching, NeRF, Gaussian Splatting, Vision-Language-Action (VLA), VAEs

AI focus

Research predictive world models using generative modeling and multimodal data to forecast scenes, support policy training and evaluation, and generalize across vehicles, robots, and sensor configurations.