Senior AI GPU Performance Engineer

About the role:

ML developers today face significant friction in taking trained models into deployment. They work in a highly fragmented space, with incomplete and patchwork solutions that require significant performance tuning and non-generalizable/ model-specific enhancements. At Modular, we are building the next generation AI platform that will radically improve the way developers build and deploy AI models.

A core part of this offering is providing a platform that allows customers to achieve state-of-the-art performance across model families and frameworks. As an AI GPU Performance Engineer, you will own architecting performance libraries for GPUs. You will develop kernels and algorithms to increase performance of kernels, reduce the activation volumes, speedup data pre- and post-processing, and improve the end-to-end performance of the AI workloads. You will collaborate with cross-functional teams to analyze and optimize AI performance on diverse GPU architectures across the stack.

LOCATION: Candidates based in the US or Canada are welcome to apply. You can work remotely from home.

What you will do:

• Design, develop, and optimize high-performance AI numeric and data manipulation kernels/operators for GPUs. 
• Achieve state-of-the-art performance by leveraging software and micro-architectural features of GPUs.
• Work with compiler, framework, runtime, and serving teams to deliver end-to-end performance that fully utilizes GPU workstations and servers. 
• Collaborate with machine learning researchers to guide system development for future ML trends.
• Engage in prototyping exercises to quantify the value proposition and develop execution plans.

What you bring to the table:

• Deep understanding of GPU computer architecture (memory hierarchies, tensor-cores, etc.) and their impact on algorithm design.
• 10+ years of relevant experience working with modern software systems and use of design patterns.
• Self motivation with the ability to design, communicate and lead new projects.
• Experience with GPU programming languages such as CUDA or OpenCL.
• Familiarity with GPU assembly such as PTX and SASS.
• Knowledge of SOTA frameworks such as CUTLASS, TRTLLM, Triton, etc.
• Able to identify industry trends, analyze emerging technologies and disruptive paradigms. 
• Creativity and curiosity for solving complex problems, a team-oriented attitude that enables you to work well with others, and alignment with our culture. 

The estimated base salary range for this role to be performed in the US, regardless of the state, is $234,000 - $319,000 USD.

The estimated base salary range for this role to be performed in Canada is $221,000 - $301,500 CAD.

The salary for the successful applicant will depend on a variety of permissible, non-discriminatory job-related factors, which include but are not limited to education, training, work experience, business needs, or market demands. This range may be modified in the future. The total compensation for a candidate will also include annual target bonus, equity, and benefits, with equity making up a significant portion of your total compensation.