Infrastructure Engineer (ML Compute)

About the role

The Infrastructure Engineer will own the GPU compute systems that power model training, evaluation, and serving. This role involves managing Kubernetes clusters, GPU hardware, and schedulers to ensure reliable production performance and efficient capacity management. You will work closely with research and product teams to optimize the underlying infrastructure stack.

Responsibilities

• Running Kubernetes clusters for GPU training and inference
• Improving how GPU work is scheduled and placed
• Detecting, draining and repairing unhealthy GPU nodes
• Managing capacity with our cloud providers
• Debugging failures in GPUs, Linux, storage and networking

Required skills

• Kubernetes
• GPU workloads
• Linux
• Python

Nice to have

• SRE
• Slurm
• Kueue
• Karpenter
• SkyPilot
• AWS EKS
• SageMaker HyperPod
• EFA
• InfiniBand
• NCCL