ML Infrastructure Engineer — Multi-Cluster GPU & Training

Prior Labs GmbH; North Canton, New York (OH, NY); On-site; USD 180,000 - 260,000; Full time

Benefits offered by this job

Remote-friendly

Offsite team events

Global offices Berlin Freiburg NewYork

Job summary

Prior Labs seeks an infrastructure engineer to own and grow its multi-cluster GPU platform. You will manage Slurm on GCP today, plan multi-provider expansion, and optimize scheduling, performance, and costs to empower researchers.

You will collaborate with researchers to understand workloads, reduce bottlenecks, and build the tooling layer and CI pipelines that keep experiments moving rapidly. Remote-optional with travel.

Qualifications

• 3+ years building and operating production GPU infrastructure at scale.
• Deep hands-on experience with Slurm and cluster management.
• Expert-level systems thinking: memory bandwidth and profiling.
• Strong Python and deep knowledge of PyTorch internals.
• Track record of improving training throughput or cost efficiency.
• Fluent AI tooling experience (Claude Code, Cursor, etc.).

Responsibilities

• Own and evolve multi-cluster GPU infrastructure across providers.
• Drive GPU utilization and training throughput; identify bottlenecks.
• Architect next-gen infrastructure: orchestration, new GPUs, capacity planning.
• Build developer productivity: CI pipelines, experiment tracking, model registry.
• Own compute budget and optimize cost per FLOP across hardware.

Skills

GPU infrastructure

Python

Systems thinking

PyTorch internals

Performance tuning

AI tooling

Tools

Slurm

GCP

Docker

wandb

GitHub Actions

Triton

Job description

Prior Labs seeks an infrastructure engineer to own and grow its multi-cluster GPU platform. You will manage Slurm on GCP today, plan multi-provider expansion, and optimize scheduling, performance, and costs to empower researchers.

You will collaborate with researchers to understand workloads, reduce bottlenecks, and build the tooling layer and CI pipelines that keep experiments moving rapidly. Remote-optional with travel.