Remote HPC GPU Cluster Engineer

ElevenLabs; United States; Hybrid; USD 140,000 - 210,000; Full time

Benefits offered by this job

Remote role - global

Office access: London, New York, SF, 1

Job summary

ElevenLabs is seeking an engineer to own and optimize our GPU infrastructure used for cutting‑edge AI research. You will operate NVIDIA GPU clusters, automate provisioning, and improve training throughput across bare‑metal and rented capacity.

You’ll partner with researchers to keep compute fast, reliable, and boring—in the best sense. This builder‑operator role offers unusual scope and autonomy within a lean team, with opportunities to push hardware and software boundaries while supporting

Qualifications

• Experience running large‑scale Linux servers or GPU environments in production.
• Deep familiarity with NVIDIA stack: drivers, CUDA, NCCL, DCGM.
• Strong scripting in Python; ability to automate end‑to‑end workflows.
• Comfort with bare‑metal hardware and high‑speed networking.
• Experience with infrastructure as code and monitoring/log analytics.

Responsibilities

• Operate and improve GPU fleet end to end: provisioning, scheduling, monitoring, upgrades, capacity planning.
• Build automation to keep the fleet healthy with minimal manual intervention.
• Own the stack beneath the training code: OS images, NVIDIA drivers, CUDA, containers.
• Run and tune job scheduling to give researchers fast, fair access.
• Maintain high‑performance storage for datasets and checkpoints.
• Identify performance issues and fix root causes, not just symptoms.
• Evaluate rented GPU capacity and ensure SLAs are met.
• Hands‑on hardware work when needed: racking, cabling, diagnostics.
• Keep clusters secure: access control, network isolation, secrets.

Skills

Linux servers

GPU clusters

Python

Automation

NVIDIA stack

Tools

Slurm

Terraform

Ansible

PromQL

Job description

ElevenLabs is seeking an engineer to own and optimize our GPU infrastructure used for cutting‑edge AI research. You will operate NVIDIA GPU clusters, automate provisioning, and improve training throughput across bare‑metal and rented capacity.

You’ll partner with researchers to keep compute fast, reliable, and boring—in the best sense. This builder‑operator role offers unusual scope and autonomy within a lean team, with opportunities to push hardware and software boundaries while supporting