over 1 year ago
Responsibilities
- Design and write high-performance, scalable software for model training
- Analyze how architectural modifications and design choices affect training throughput and quality
- Write low-level CUDA and Triton kernels to optimize accelerator performance
- Research, implement, and experiment with ideas across supercompute and data infrastructure
- Develop performance and profiling tools to identify and remove bottlenecks
- Collaborate with researchers on efficient and reliable language understanding and generation systems
Requirements
- Extremely strong software engineering skills
- Proficiency in Python and related machine-learning frameworks including JAX, PyTorch, and XLA/MLIR
- Experience writing GPU kernels using CUDA, Triton, or similar technologies
- Experience with large-scale distributed training strategies
- Familiarity with autoregressive sequence models such as Transformers
- A paper at a top-tier venue such as NeurIPS, ICML, ICLR, AIStats, MLSys, JMLR, AAAI, Nature, COLING, ACL, or EMNLP is a bonus
Benefits
- Remote-flexible work with offices in Toronto, New York, San Francisco, London, and Paris, plus a co-working stipend
- Weekly lunch stipend, in-office lunches, and snacks
- Full health and dental benefits, including a separate mental health budget
- 100% parental leave top-up for up to 6 months
- Personal enrichment benefits for arts and culture, fitness and well-being, quality time, and workspace improvement
- Six weeks of vacation, or 30 working days
- Inclusive work environment and reasonable accommodations during recruitment
Categories
About Cohere
Cohere is the leading security-first enterprise AI company. We build cutting-edge foundation models and end-to-end AI products designed to solve real-world business problems. We partner closely with companies to deliver seamless integration, full customization, and easy-to-use solutions for their workforce and customers. Our all-in-one platform offers enterprises the highest levels of data security, privacy and optionality to deploy across all major cloud providers, private cloud environments, or on-premises. HQ: 171 John Street, 2nd Floor, Toronto, ON M5T 1X3