PhD AI Research Intern: World Models & Multimodal Gen

Xpengmotors; Santa Clara (CA); On-site; USD 28,000 - 46,000; Full time

Benefits offered by this job

Competitive compensation

Snacks, lunches, dinners

Job summary

XPENG seeks a PhD research intern to advance world models and generative architectures for autonomous driving and robotics. You will work with a research team, design focused projects, and publish results where appropriate.

You will explore high-quality multi-view future prediction, 3D scene representations, and cross-embodiment generalization across vehicles and sensors, with access to vast multimodal data from XPENG’s fleet and platforms.

Qualifications

• Currently pursuing a PhD in Engineering, Computer Science, or a related field with a focus on DL, CV, or Generative Models.
• First-author publications at top-tier venues; arXiv preprints acceptable.
• Strong, up-to-date foundation in generative modeling and experimental methodology.
• Strong Python programming and software design, with experience in large-scale data processing.

Responsibilities

• Drive a focused research project on predictive world models from problem formulation to evaluation.
• Contribute to multi-view future prediction and action-conditioned or unconditioned forecasting.
• Extend prediction into a multimodal latent space with 3D scene representations.
• Investigate cross-embodiment generalization across vehicles, robots, and sensors.
• Build evaluation methodology for prediction quality, fidelity, and policy performance.
• Collaborate with researchers to move prototypes into scalable training and publish results.

Skills

Deep Learning

Computer Vision

Generative Modeling

PyTorch

Python

JAX

Research publications

Education

PhD in Engineering/CS

Tools

PyTorch

JAX

Git

Job description

XPENG seeks a PhD research intern to advance world models and generative architectures for autonomous driving and robotics. You will work with a research team, design focused projects, and publish results where appropriate.

You will explore high-quality multi-view future prediction, 3D scene representations, and cross-embodiment generalization across vehicles and sensors, with access to vast multimodal data from XPENG’s fleet and platforms.