Anthropic’s TPU Kernel Engineer role focuses on improving the performance of machine learning systems on TPUs and other accelerators. The engineer will analyze performance bottlenecks, design efficient kernels, and collaborate closely with researchers to ensure model changes do not degrade system throughput. The position is hybrid, requiring presence in one of the U.S. offices at least 25% of the time, and offers a competitive salary range, equity options, and comprehensive benefits.
TPU Kernel Engineer at Anthropic
Hybrid - San Francisco, CA, USA; New York City, NY, USA; Seattle, WA, USA
More jobs at AnthropicSalary
USD 280,000 - 850,000
Requirements
Skills
- Significant experience optimizing ML systems for TPUs, GPUs, or other accelerators
- Results-oriented, with a bias towards flexibility and impact
- Ability to pick up slack, even if it goes outside the job description
- Enjoyment of pair programming
- Interest in learning more about machine learning research
- Concern for societal impacts of work
- High performance, large-scale ML systems experience
- Experience designing and implementing kernels for TPUs or other ML accelerators
- Background in computer architecture or deep understanding of accelerators
- Knowledge of ML framework internals
- Experience with language modeling and transformers
- Bachelor’s degree or an equivalent combination of education, training, and/or experience
Responsibilities
- Identify and address performance issues across many different ML systems, including research, training, and inference
- Design and optimize kernels for the TPU
- Provide feedback to researchers about how model changes impact performance
- Solve large-scale systems problems and low-level optimization challenges
Technologies
TPUGPUML systemsLow-level optimizationKernel design and implementationML framework internalsTransformersLow-precision inferenceCustom collective communication algorithms
See if your resume is ready for this job
See how our AI can optimize your resume and improve your chances for this role.