PhD AI Research Intern: World Models for Driving

XPENG; Santa Clara (CA); On-site; USD 70,000 - 110,000; Full time

Benefits offered by this job

Competitive compensation

Snacks & meals

On‑site collaboration

Job summary

XPENG seeks PhD research interns in Santa Clara to advance world models that learn dynamics from large multimodal data. You will study diffusion/flow-matching models, video tokenizers, and transformer backbones on autonomous driving data, with an emphasis on end‑to‑end research projects suitable for top‑tier publications.

Ideal candidates are pursuing a PhD in engineering or CS, have publications, and are skilled in PyTorch or JAX.

Qualifications

• Currently pursuing a PhD in Engineering, Computer Science, or a related field, with a focus on Deep Learning, Computer Vision, or Generative Models.
• First‑author publications at top‑tier venues (CVPR, ICCV, ECCV, NeurIPS, ICLR, ICML, CoRL, RSS, SIGGRAPH) or arXiv submission.
• Hands‑on experience building, training, fine‑tuning, and evaluating models in PyTorch or JAX.

Responsibilities

• Drive a focused research project on predictive world models, including problem formulation, architecture design, training, evaluation, and empirical analysis.
• Contribute to multi‑view future prediction, action conditioned rollouts, and multimodal backbones.
• Extend prediction into a 3D multimodal latent space with 3D scene representations and rewards signals.
• Investigate cross‑embodiment generalization across vehicles/robots with few‑shot data.
• Build evaluation methodologies for representation quality, accuracy, fidelity, and long‑horizon rollout consistency.
• Collaborate with research engineers to move prototypes into scalable training and publish/open‑source results.

Skills

Python programming

PyTorch

JAX

Experimental methodology

Education

PhD candidate in Engineering/CS

Tools

PyTorch

JAX

Job description

XPENG seeks PhD research interns in Santa Clara to advance world models that learn dynamics from large multimodal data. You will study diffusion/flow-matching models, video tokenizers, and transformer backbones on autonomous driving data, with an emphasis on end‑to‑end research projects suitable for top‑tier publications.

Ideal candidates are pursuing a PhD in engineering or CS, have publications, and are skilled in PyTorch or JAX.