GPU Infrastructure Engineer — Remote & Multi-Cloud Clusters

Cohere; New York (NY); Hybrid; USD 140,000 - 210,000; Full time

Benefits offered by this job

Weekly lunch stipend

Full health and dental benefits

RRSP matching / 401K / Pension

6 weeks paid vacation

Parental leave top-up

Education stipend

Office travel budget

Home office stipend $500

Job summary

Cohere seeks an experienced Infrastructure Engineer to manage GPU Kubernetes clusters across multi-cloud environments and support ML workloads. You will work with AI researchers and ML engineers to ensure stability, scalability, and observability of Cohere's infrastructure.

The role involves automating fault detection, optimizing fleet utilization, and reducing downtime. We value collaboration with researchers and partners, and you will contribute to open-source projects where applicable.

Qualifications

• Run GPU Kubernetes clusters for ML/HPC workloads.
• Debug Linux networking and performance issues; RDMA/NCCL knowledge is a plus.
• Write code in Python or Go; contribute to open-source projects.

Responsibilities

• Deploy and operate GPU Kubernetes clusters across clouds.
• Collaborate with cloud providers and hardware partners on reliability and cost.
• Automate node, GPU, and fabric fault detection and remediation.

Skills

GPU Kubernetes

RDMA

NCCL

Python

Go

Open-source

Tools

Kubernetes

Job description

Cohere seeks an experienced Infrastructure Engineer to manage GPU Kubernetes clusters across multi-cloud environments and support ML workloads. You will work with AI researchers and ML engineers to ensure stability, scalability, and observability of Cohere's infrastructure.

The role involves automating fault detection, optimizing fleet utilization, and reducing downtime. We value collaboration with researchers and partners, and you will contribute to open-source projects where applicable.