Remote-First AI Factory Network Architect (HPC Fabric Lead)

Runware; United States; Remote; USD 180,000 - 240,000; Full time

Benefits offered by this job

Generous paid time off

Stock options

Flexible hours

Family leave

Company retreats

Job summary

Runware is building the AI supercomputer fabric across US and Europe. You will own the end-to-end network design from server NVLink inside machines to pod-wide interconnects, shaping a scalable, high-performance fabric and leading a growing team of network engineers.

This hands-on leadership role requires deep experience with InfiniBand, NVLink, optical networks, and RDMA tuning. You’ll work with NVIDIA, optics vendors, and site teams to deliver a reliable, scalable, and observable network.

Qualifications

• 15+ years designing, building, and operating high-performance networks.
• Deep hands-on InfiniBand fabric knowledge and routing experience.
• Experience with NVLink/NVSwitch and multi-GPU interconnects.
• Strong automation and telemetry capability for large-scale fabrics.
• Willingness to travel to pod sites in US and Europe.

Responsibilities

• Own end-to-end network architecture for the AI Factory fabric including scale-up and scale-out networks.
• Design the optical layer and inter-site links with reliability targets.
• Define and validate repeatable pod fabric designs for fast bring-up.
• Build and bring up initial fabrics, configure routing, QoS, and RDMA tuning.
• Collaborate with engineering teams and vendors to drive topology decisions.
• Lead capacity planning and vendor evaluations for new fabric generations.
• Set standards for telemetry, health monitoring, and runbooks for tens of thousands of ports.
• Hire and lead network engineers as the footprint grows and provide escalation support.

Skills

InfiniBand

NVLink

NVSwitch

RDMA

GPUDirect

NCCL

Python

Go

Telemetry

Automation

Tools

NVIDIA Spectrum-X

Quantum-X800

Blackwell NVLink

OpenFabrics

SONiC

Job description

Runware is building the AI supercomputer fabric across US and Europe. You will own the end-to-end network design from server NVLink inside machines to pod-wide interconnects, shaping a scalable, high-performance fabric and leading a growing team of network engineers.

This hands-on leadership role requires deep experience with InfiniBand, NVLink, optical networks, and RDMA tuning. You’ll work with NVIDIA, optics vendors, and site teams to deliver a reliable, scalable, and observable network.