About the role#
This research internship focuses on the development of interactive Video World Models. The goal is to build foundational models that understand physics, causality, and complex dynamics to simulate virtual worlds. This is a technical role that bridges generative AI research with high-performance machine learning systems, requiring both architectural design and infrastructure engineering.
What you'll do#
- Design, train, and scale interactive world models using techniques such as Diffusion, Autoregressive Transformers, VAEs, LLMs, and VLMs.
- Architect highly scalable distributed training pipelines, implementing advanced model and data parallelism for large-scale computing clusters.
- Profile and optimize model architectures to address memory and compute bottlenecks.
- Write high-performance, custom hardware kernels to maximize Model FLOPs Utilization (MFU) and support low-latency inference.
What you'll need#
- Currently pursuing a PhD or a Master’s degree in Computer Science, Machine Learning, Computer Architecture, or a related field.
- Exceptional coding proficiency in Python or other languages.
- Experience with modern AI infrastructure stacks and large-scale machine learning systems, such as PyTorch FSDP or Megatron.
- Thorough understanding of modern generative paradigms, including Diffusion, Vision Transformers, and discrete tokenization.
- First-author publications in top-tier AI or ML Systems venues (e.g., NeurIPS, ICLR, ICML, CVPR, ICCV, MLSys, OSDI, ASPLOS).
Location & details#
- Location: Palo Alto, California.
- Modality: On-site.
- Employment: Full-time.
- Compensation: The base pay range is $80,168.40 to $124,800.00 per year.
- Benefits: Eligible for paid sick leave, up to 13 paid holidays, and enrollment in the company-sponsored medical plan.
