Staff Training Infra Engineer for Scalable ML Pipelines

Cohere; Paris; Hybrid; EUR 90,000 - 150,000; Full time

Benefits offered by this job

Lunch stipend & meal benefits

Health and dental benefits

RRSP matching / Pension plan

Parental leave (6 months)

Learning stipend & conferences

Vacation 6 weeks / 30 days

Office budget for offsite travel

Home office stipend $500

Job summary

Cohere is seeking a highly skilled Member of Technical Staff to contribute to training pipelines, ship state-of-the-art models, and bridge research with production. The role emphasizes writing scalable training software, enhancing infrastructure performance, and collaborating with top researchers.

Offices are available globally with remote-friendly options, and a strong emphasis is placed on advancing compute-to-engineering balance.

Qualifications

• Strong software engineering skills and ability to write production-grade code.
• Proficiency in Python and ML frameworks (JAX, PyTorch, XLA/MLIR).
• Experience with distributed training infrastructures and large-scale training setups.

Responsibilities

• Design and write high-performant and scalable software for training.
• Improve training setup from infra and codebase performance standpoint.
• Craft and implement tools to speed up training cycles and improve infrastructure efficacy.
• Research, implement, and experiment with ideas on supercompute and data infrastructure.
• Learn from and work with leading researchers in the field.

Skills

Software engineering

Python & ML frameworks

Distributed training infra

Large-scale training

Research publications

Tools

Kubernetes

Slurm

Ray

JAX

PyTorch

XLA/MLIR

Job description

Cohere is seeking a highly skilled Member of Technical Staff to contribute to training pipelines, ship state-of-the-art models, and bridge research with production. The role emphasizes writing scalable training software, enhancing infrastructure performance, and collaborating with top researchers.

Offices are available globally with remote-friendly options, and a strong emphasis is placed on advancing compute-to-engineering balance.