Staff ML Engineer - Training & Infra

Visa Hunt; Paris; Sur place; EUR 131 845 - 184 583; Plein temps

Avantages offerts par ce poste

Remote-friendly

Home office stipend

Annual offsite

Healthcare benefits

Résumé du poste

Cohere is seeking a Staff Software Engineer to design and optimize training pipelines for large AI models. You will contribute to scalable software and infrastructure that bridges research and production, with access to substantial compute and data resources.

The role welcomes remote work with global offices and a collaborative culture. You will work closely with researchers to push state-of-the-art models into production, expanding tooling and training workflows.

Qualifications

• Extremely strong software engineering skills.
• Proficiency in Python and ML frameworks such as JAX and PyTorch.
• Experience with distributed training infrastructures (Kubernetes, Slurm) and frameworks (Ray).
• Experience with large-scale distributed training strategies.
• Hands-on experience training large models and building tooling for training infra.

Responsabilités

• Design and write high-performant, scalable software for training.
• Improve training setup from infra and codebase perspective.
• Craft and implement tools to speed up training cycles and improve infra efficacy.
• Research and experiment with ideas on supercompute and data infra.
• Collaborate with top researchers to advance production-readiness.

Connaissances

Strong software engineering

Python

ML frameworks (JAX, PyTorch)

Distributed training (Kubernetes, Slur

Large-scale training

Research collaboration

Outils

Kubernetes

Slurm

Ray

JAX

PyTorch

Description du poste

Cohere is seeking a Staff Software Engineer to design and optimize training pipelines for large AI models. You will contribute to scalable software and infrastructure that bridges research and production, with access to substantial compute and data resources.

The role welcomes remote work with global offices and a collaborative culture. You will work closely with researchers to push state-of-the-art models into production, expanding tooling and training workflows.