PhD Research Intern, Hardware and Systems Architecture - 2027

NVIDIA•Us, CA

About The Position

NVIDIA's development of the GPU in 1999 revolutionized the PC gaming industry, reshaped contemporary computer graphics, and transformed parallel computing. In recent years, GPU deep learning has propelled modern AI, ushering in a new era of computing where GPU systems serve as the core intelligence for computers, robots, and autonomous vehicles, enabling them to perceive and interpret the world around them. NVIDIA Research is at the forefront of developing cutting edge hardware, software, and algorithmic technologies for next generation GPU systems. We are eager to expand our teams with the most brilliant minds globally.

Requirements

Pursuing PhD Degree in relevant discipline(s) (CS, CE, EE, Physics, Math)
At least one year of university or industrial research experience in computer architecture or computer systems, including operating systems, distributed systems, compilers, or AI/ML systems.
Strong programming ability: C, C++ and Python preferred.
Eagerness to learn and use AI tools for software development and system design.
Experience designing and evaluating research ideas through prototyping, simulation, modeling, or experimentation.
Strong programming skills and the ability to communicate research ideas, methods, and results clearly.

Nice To Haves

Chip/system level arch: agentic techniques for computer architecture; specialized accelerators for deep-learning and humanoid robotics
Systems for AI/ML: systems infrastructure for LLM training, inference, and agents/tools; state management in agentic systems; systems for humanoid robotics
AI/systems co-design
Experience contributing to research targeting leading architecture and systems conferences, such as ISCA, ASPLOS, MICRO, HPCA, SOSP, MLSys, NSDI, or OSDI.
Publications are a plus but are not required.
Candidates with research in progress, including manuscripts in preparation or submissions under review.

Responsibilities

Conduct cutting edge research on algorithms, hardware, and system software extending the state of the art in deep learning, humanoid robotics, database analystics, general computing, and other technology areas surrounding NVIDIA's business.
Invent new techniques, technologies, methodologies, processes, and devices that can enable new products or types of products.
Deliverable results include prototypes, patents, leading to products and publications.
Develop new methods and processes for research and scientific evaluation of research techniques and artifacts including agentic driven techniques, simulation, modeling, and benchmarking.
Produce technology vision and the basis for products 5-10 years out. Focus should not be on products currently shipping or in development, except as to how they can be extended and improved.