New
Senior Researcher
![]() | |
![]() United States, Washington, Redmond | |
![]() | |
OverviewArtificial Intelligence (AI)-assisted software generation has rapidly progressed, but a significant gap remains in enabling AI to generate high-performance GPU/NPU kernels. We are looking for a Senior Researcher to help close this gap. This opportunity is ideal for individuals with expertise in modern AI and machine learning techniques who are passionate about advancing AI-for-AI systems.This work is central to Microsoft's next-generation AI infrastructure. The innovations developed in this role will operate across thousands of GPUs and MAIA AI accelerators in production, shaping how Microsoft trains and serves frontier AI systems at global scale. It's a unique opportunity to make foundational contributions that influence Microsoft's AI stack end to end-from research and kernel generation to deployment in hyperscale environments.
Responsibilities Apply large language models (LLMs) to generate, optimize, and refactor compute kernels for performance and portability across a wide range of hardware platforms.Leverage cutting-edge techniques-including reinforcement learning with human and automated feedback (RLHF, RLAF), test-time scaling, and agentic workflows-to produce high-performance kernels.Explore opportunities to replace traditional compiler algorithms with AI-driven approaches that improve performance, maintainability, and adaptability.Develop and implement intelligent strategies for kernel fusion that reduce memory bandwidth demands and enhance runtime efficiency.Partner with compiler architects, machine learning researchers, and systems engineers to embed AI capabilities into the compiler toolchain.Stay at the forefront of AI and compiler advancements by rapidly prototyping and evaluating new ideas in a collaborative and inclusive environment.Embody Microsoft's Culture and Values |