Member of Technical Staff, Training Performance Engineer

About Cohere

Cohere is a pioneering enterprise AI company prioritizing security. We develop innovative foundational AI models and comprehensive products aimed at addressing practical business challenges. Our team consists of dedicated researchers, engineers, and designers passionate about advancing AI technology to drive value for clients globally. Headquartered in Toronto with multiple international offices, we foster inclusive collaboration worldwide.

Role Summary

The Training Performance Engineer will focus on enhancing the efficiency of large-scale language model training within the Pre-Training team. This position entails optimizing training throughput and maximizing accelerator hardware utilization through sophisticated software engineering and kernel-level development. Working at the intersection of system design and machine learning, you will identify performance constraints, build advanced profiling tools, and contribute to the innovation of natural language processing frameworks.

Responsibilities

• Develop scalable, high-performance software tailored to training large-scale language models.
• Analyze how architectural decisions influence training speed and model quality.
• Implement and optimize low-level CUDA and Triton kernels to maximize accelerator efficiency.
• Conduct research experiments on supercomputing and data platforms.
• Collaborate with leading domain researchers to enhance training systems.

Qualifications

• Outstanding skills in software engineering.
• Expertise in Python alongside ML frameworks like JAX, PyTorch, and XLA/MLIR.
• Proven experience writing GPU kernels with CUDA and Triton.
• Knowledge of distributed training techniques across large systems.
• Familiarity with transformer-based autoregressive sequence models.

Additional Information

This position supports remote work aligned primarily with ET to CET time zones for optimal team synergy. Candidates are encouraged regardless of their office proximity. Cohere promotes a diverse and inclusive workplace welcoming applications from all backgrounds and provides accommodation support during recruitment.

Employee Benefits

• Weekly lunch allowance worth $75 (or equivalent local currency).
• Comprehensive health and dental coverage including mental health resources.
• Retirement savings plans with matching contributions (RRSP, 401K, Pension).
• Full parental leave top-up for six months available to either parent.
• Annual enrichment funds for culture, wellness, workspace, and education.
• Generous vacation policy offering six weeks (30 working days) of paid leave.
• Travel budget supporting visits to other offices and company-wide offsites.

Work Environment

• Remote-friendly company with offices worldwide including Toronto, London, New York, San Francisco, Montreal, Paris, Berlin, and Seoul.
• Workspaces equipped with daily lunch, snacks, and social opportunities for on-site employees.
• Co-working space allowance for remote employees.
• $500 stipend for setting up a proper home office.

Compensation

The salary range for this role is from CA$250,000 to CA$535,000 annually.

Note on Application and Security

Cohere does not charge candidates for any hiring processes or external services. Official communications come exclusively from the company domain. AI tools may be used to assist in application review but do not limit recruiter evaluations.