Advance model post-training with Python and large-scale distributed training at Cohere, working remotely or from listed offices. Ship frontier models into production.
As a Member of Technical Staff, you will bridge research and production while advancing model post-training. You will write production software, support research, and collaborate across specialist teams. The work spans SFT, RL, supercompute, and data infrastructure.
Key Responsibilities: Design scalable software for model training
Post-train models toward state-of-the-art performance
Coordinate with Agentic and Code specialist teams
Improve SFT and RL training cycles
Experiment on supercompute and data infrastructure
Requirements: Extremely robust software engineering skills
Proficiency in Python, JAX, PyTorch, and XLA/MLIR
Experience with Kubernetes, Slurm, and Ray
Experience using large-scale distributed training strategies
Hands-on experience training large models at scale
Build production machine learning systems while optimizing post-training performance across Cohere’s research and engineering efforts.