TPU Performance Engineer

About the company

Anthropic's mission is to create reliable, interpretable, and steerable AI systems that are safe and beneficial for users and society.

Responsibilities

• Identify and address performance issues across ML systems, including research, training, and inference
• Design and optimize kernels for the TPU
• Provide feedback to researchers about how model changes impact performance
• Implement low-latency, high-throughput sampling for large language models
• Adapt existing models for low-precision inference
• Build quantitative models of system performance
• Design and implement custom collective communication algorithms
• Debug kernel performance at the assembly level

Requirements

• Significant experience optimizing ML systems for TPUs, GPUs, or other accelerators
• Track record of solving large-scale systems problems and low-level optimization
• Experience with high-performance, large-scale ML systems
• Understanding accelerators at a deep level, e.g. background in computer architecture
• Familiarity with ML framework internals
• Language modeling with transformers
• Bachelor's degree or equivalent combination of education, training, and/or experience

Conditions

• Annual salary: $280,000 - $850,000 USD
• Location-based hybrid policy: at least 25% of time in office
• Visa sponsorship available
• Optional equity donation matching, generous vacation and parental leave, flexible working hours