About the role#
The Deep Learning Efficiency Research team is looking for a PhD intern to focus on efficient deep learning methods. You will work on research that aims to have real-world impact. Our core areas include efficient diffusion language models, multimodal generative models, and agentic AI with hybrid inference orchestration. You will have the opportunity to contribute to projects that may be integrated into products.
What you'll do#
- Research, design, and implement new methods for efficient deep learning.
- Develop diffusion LLMs and multimodal models, focusing on areas like sampling efficiency, adaptive unmasking, self-speculation, and distillation pipelines.
- Build efficient agentic AI systems, including hybrid inference orchestration across cloud and edge, routing and scheduling policies, and resource-aware agent loops.
- Publish original research findings.
- Collaborate with internal team members, product groups, and external researchers to transfer technology.
What you'll need#
- Current enrollment in a PhD program for Computer Science, Computer Engineering, Electrical Engineering, or a related field.
- Strong knowledge of the theory and practice of machine learning and deep learning.
- Experience with large language models, diffusion language models, multimodal models, or agentic systems.
- Hands-on experience with large-scale model training, specifically data preparation and model parallelization techniques like tensor and pipeline parallelism.
- A research track record with at least one publication at a top-tier conference such as ICML, ICLR, NeurIPS, CVPR, or ICCV.
- Excellent communication skills.
Location & details#
- This is a paid, full-time internship.
- The role is available as both remote and on-site in Santa Clara, California.
- The position follows a rolling term.
- Please note that visa sponsorship is not provided for this role.
About NVIDIA
NVIDIA operates as a computer hardware manufacturer based in Santa Clara, California. Founded in 1993, the company focuses on accelerated computing and graphics technology. It produces hardware for markets including artificial intelligence, gaming, and data centers. The organization employs over 50,000 people and maintains a global presence.
How to get in at NVIDIA
Applying early gives you a distinct advantage at NVIDIA because recruiters review applications as they arrive. Intern Insider sends an instant alert the moment a role matching your target is published, so you can apply among the first before the pile grows. Getting your resume in front of a human early is often the difference between a screening call and a rejection. You can use Intern Insider to surface the recruiters behind NVIDIA roles to reach out directly. Asking a recruiter about the role or a referral materially improves your response rates compared to submitting into a general queue. It is a simple way to make your application stand out in a competitive process.
