Cohere is seeking a Performance Engineer in Greater London to optimize the performance of advanced language models and systems. You will focus on improving model training metrics and ensuring high accelerator utilization.

The role involves designing scalable software, writing low-level CUDA kernels, and experimenting with ideas on super-compute infrastructure. Applicants should have strong software engineering skills and proficiency in Python, with experience in GPUs and distributed training.

Cohere supports a remote-friendly work environment, offering benefits like co-working spaces for remote teams.