Remote Senior GPU Network Engineer – Fabric & Cluster Builds

REALM; United States; On-site; USD 170,000 - 230,000; Full time

Job summary

Realm is hiring a senior network architect to own the GPU training fabric at scale. You’ll design buildouts, lead deployments, and act as the main escalation point for performance incidents, enabling high-throughput AI workloads.

You’ll collaborate with global customers, automate repetitive tasks with Python and Ansible, and mentor teammates. This is a fully remote role within the United States with flexible scheduling for maintenance windows.

Qualifications

• 7+ years in network operations or engineering with growing responsibilities.
• Hands-on GPU cluster networking: RDMA/RoCEv2 fabric config and lossless Ethernet design.
• Expert-level BGP on internet-scale networks and ECMP awareness.
• Strong Linux networking know-how: kernel routing, nftables/iptables, FRR/Bird.
• Comfort with multiple vendors/NOS: Juniper, Cisco, Nokia, Cumulus, SONiC.
• Proficiency in Python and Ansible for repeatable changes and data collection.

Responsibilities

• Own the fabric behind GPU clusters; diagnose latency issues and fix root causes.
• Lead design and deployment of GPU cluster buildouts and migrations globally.
• Serve as senior escalation for critical network incidents with cross-functional coordination.
• Onboard strategic customers to GPU deployments and resolve performance problems.
• Champion automation; develop tooling to reduce manual toil.
• Mentor engineers and raise the team's baseline through docs and pairing.
• Participate in 24x7x365 on-call rotation and incident leads.

Skills

Networking

GPU networking

BGP

Linux networking

Python

Ansible

Automation

Education

Bachelor's degree in a relevant field

Tools

Juniper

Cisco

Nokia

Cumulus

SONiC

nftables

iptables

NVLink/NVSwitch

RDMA/RoCEv2

Job description

Realm is hiring a senior network architect to own the GPU training fabric at scale. You’ll design buildouts, lead deployments, and act as the main escalation point for performance incidents, enabling high-throughput AI workloads.

You’ll collaborate with global customers, automate repetitive tasks with Python and Ansible, and mentor teammates. This is a fully remote role within the United States with flexible scheduling for maintenance windows.