# Member of Technical Staff, Training Performance Engineer

Hiring organization: [Cohere](https://career.thegoodapps.co/organizations/cohere)

Canonical page: https://career.thegoodapps.co/jobs/9d4cfb7a-44ce-4be5-8975-7ecd2b67eccd

Listed on Cohere's own careers site. Applications go to them directly.

- Employment type: full time
- Location: New York, NY
- Remote: yes
- Salary: 250000 – 535000 CAD per year

## Summary

This role optimizes the training performance of large language models by eliminating bottlenecks and writing low-level GPU kernels. It suits engineers with strong systems programming skills who want to work on the infrastructure that makes frontier AI models efficient.

_Our summary, not Cohere's wording._

## Skills named

CUDA, JAX, Python, PyTorch

## Required

- Extremely strong software engineering skills
- Proficiency in Python and ML frameworks (JAX, PyTorch, XLA/MLIR)
- Experience writing GPU kernels with CUDA or Triton
- Experience with large-scale distributed training
- Familiarity with autoregressive sequence models like Transformers

## Nice to have

- Published papers at top-tier ML venues (NeurIPS, ICML, ICLR, AIStats, MLSys, JMLR, AAAI, Nature, COLING, ACL, EMNLP)

Apply on Cohere's site: https://jobs.ashbyhq.com/cohere/d42f5fd4-1ffc-45b9-957c-f09862db6af6/application