ML Infra Engineer: GPU Fleet & Inference Orchestration

Generalist; Somerville (NJ); On-site; USD 140,000 - 190,000; Full time

Job summary

Generalist builds large-scale robot foundation models and relies on massive GPU fleets to train and run experiments. You will own the GPU compute fleets and ensure researchers can run distributed training and real-time inference on demanding hardware.

You will optimize data loading, storage, and network pipelines across distributed environments and orchestrate robot inference fleets with Slurm or Kubernetes.

Qualifications

• Experience managing large GPU fleets for distributed training and inference.
• Strong knowledge of ML data loading and storage.
• Proficiency with NVIDIA GPU ecosystem and hardware/software stack.

Responsibilities

• Own our GPU compute fleets.
• Ensure GPUs are easy for researchers to use and maximally utilized.
• Optimize ML data loading transport and storage in distributed environments.
• Orchestrate robot inference fleets.

Skills

GPU orchestration

Distributed training

ML data loading

Real-time inference

NVidia GPU ecosystem

Tools

Slurm

Kubernetes

Job description

Generalist builds large-scale robot foundation models and relies on massive GPU fleets to train and run experiments. You will own the GPU compute fleets and ensure researchers can run distributed training and real-time inference on demanding hardware.

You will optimize data loading, storage, and network pipelines across distributed environments and orchestrate robot inference fleets with Slurm or Kubernetes.