Senior HPC Systems Engineer: GPU/InfiniBand & KVM

Nebius; United States; On-site; USD 170,000 - 300,000; Full time

Benefits offered by this job

Competitive compensation

Career growth

Flexible and ownership culture

Opportunity to work on impactful AI

Job summary

Nebius is seeking a Senior Software Systems Engineer to advance our hyperscaler platform. You will work on GPU computing, InfiniBand networks, and the KVM/QEMU stack, improving performance and security in multi‑GPU HPC environments.

Role involves analyzing, troubleshooting, and automating fault detection within a complex system, with a focus on hardware virtualization and device emulation. Collaboration with hardware teams is essential.

Qualifications

• 5+ years of professional experience in system‑level software development with focus on performance optimization.
• 3+ years of hands‑on Linux system administration and tuning.
• Deep understanding of server architecture, PCIe devices, NICs, Linux OS/Kernel and HPC systems.
• Proficiency in C/C++, Go or Python for performance‑oriented programming.
• Experience with GPU/InfiniBand in clusters and virtualization stacks (KVM/QEMU).

Responsibilities

• Tune GPU clusters and InfiniBand networks for high‑performance operation in HPC environments.
• Analyze and troubleshoot GPU/InfiniBand issues, propose corrective actions.
• Integrate new hardware into existing infra, including GPU support via Kubernetes, QEMU and KVM.
• Enhance automation for monitoring and proactive fault detection in GPU/InfiniBand setups.
• Configure and manage GPU devices and InfiniBand fabrics for reliability and performance.

Tools

Kubernetes

QEMU

KVM

Job description

Nebius is seeking a Senior Software Systems Engineer to advance our hyperscaler platform. You will work on GPU computing, InfiniBand networks, and the KVM/QEMU stack, improving performance and security in multi‑GPU HPC environments.

Role involves analyzing, troubleshooting, and automating fault detection within a complex system, with a focus on hardware virtualization and device emulation. Collaboration with hardware teams is essential.