Lead GPU Cluster Architect - AI Infra (Remote)

Praxy; Northern (KY); Hybrid; USD 184,000 - 318,000; Full time

Benefits offered by this job

Health insurance

401(k) plan

Parental leave

Remote work reimbursement

Job summary

Nebius is seeking a GPU Cluster Architect to drive end-to-end design of scalable AI infrastructure and GPU interconnectivity. This remote US-based role involves defining architectures for tens of thousands of GPUs across data center sites, balancing performance, reliability, and power efficiency.

You will collaborate with SRE, networking, storage, and DC teams to deliver high-impact AI compute platforms and optimized data paths.

Qualifications

• 5+ years designing GPU clusters.
• Deep understanding of modern GPU architectures (NVIDIA, AMD).
• Experience with InfiniBand and RoCE interconnects.
• Strong background in systems architecture, networking, and hardware reliability.
• Scripting for automation and telemetry (Python, Go).

Responsibilities

• Cluster Design: architect scalable GPU cluster topologies (compute nodes, interconnects, storage, control planes).
• Performance Modeling: analyze AI/ML workloads to inform design tradeoffs (latency, bandwidth, GPU density).
• Network Architecture: validate low-latency, high-throughput interconnects at data-center scale.
• Storage Integration: optimize performance for training datasets, checkpoints.
• Reliability & Monitoring: interpret monitoring signals to detect design gaps.
• Collaboration: work with site reliability, networking, storage, and DC engineering teams.

Skills

GPU architecture

HPC interconnects

Systems architecture

Scripting (Python/Go)

Cluster design

Tools

Python

Go

Job description

Nebius is seeking a GPU Cluster Architect to drive end-to-end design of scalable AI infrastructure and GPU interconnectivity. This remote US-based role involves defining architectures for tens of thousands of GPUs across data center sites, balancing performance, reliability, and power efficiency.

You will collaborate with SRE, networking, storage, and DC teams to deliver high-impact AI compute platforms and optimized data paths.