Summary

This infrastructure engineer will operate GPU compute systems supporting model training, evaluation, and serving. The role spans Kubernetes clusters, GPU scheduling, node health, and cloud capacity, with a hybrid schedule of three days per week in the Paris or London office and occasional travel between them.

Responsibilities

• Run Kubernetes clusters for GPU training and inference
• Improve GPU workload scheduling and placement
• Detect, drain, and repair unhealthy GPU nodes
• Manage compute capacity with cloud providers
• Debug failures across Linux, storage, and networking