Member of Technical Staff, Training Performance Engineer
Cohere · London · New York · Paris · Toronto · Montreal
CHECKED 44M AGO
Frontier labremote$175k–$375k
At a glance
Highlights
- Optimize performance of frontier language models in pre-training.
- Remove bottlenecks to boost training throughput and accelerator use.
- Build cutting-edge training and profiling tools for large models.
What they offer
- Equity
- Full health and dental benefits, including mental health budget
- 6 weeks paid vacation
- 100% parental leave top-up for up to 6 months
- $500 home office stipend and co-working benefit
Ideal candidate
A strong software engineer passionate about performance and large-scale ML who loves low-level optimization and collaborating with researchers to push frontier model efficiency.
Requirements
- Extremely strong software engineering skills
- Proficiency in Python and related ML frameworks such as JAX, PyTorch and XLA/MLIR
- Experience writing kernels for GPUs using CUDA, Triton, etc.
- Experience using large-scale distributed training strategies
- Familiarity with autoregressive sequence models, such as Transformers
- remote policy
- remote-friendly
Lab profile
About Cohere
Security-first enterprise AI company building frontier foundation models for business applications.
What they build
- Command foundation models
- Embed models
- Rerank models
Known for
- Security-first enterprise AI
- Frontier model training and deployment
- Business-focused end-to-end products
Key facts
- Founded in 2019
- Headquartered in Toronto, Canada