Lead AI Factory Network Architect — Remote

JoinStartup; United States; Remote; USD 180,000 - 260,000; Full time

Job summary

Runware seeks an AI Factory Network Architect to own the design, build, and operation of the global AI supercomputer network. You will lead hands‑on fabric bring‑up, set standards, and collaborate with hardware, software, and vendor teams across US and Europe.

You will drive scale‑out and scale‑up fabrics using NVLink/NVSwitch, InfiniBand, RoCE, and optical links, while ensuring fault‑tolerance, telemetry, and incident runbooks for tens of thousands of ports.

Qualifications

• 15+ years designing, building, and operating high‑performance networks in HPC/AI scale
• Deep hands‑on InfiniBand fabric design, subnet management, and routing at port level
• Experience with NVLink/NVSwitch for scale‑up and scale‑out fabrics
• Optical networking depth: DWDM, transceivers, and inter‑site links at scale
• Hands‑on RDMA across InfiniBand and RoCE; GPUDirect and PCCL/NCCL considerations
• Automation: Python/Go, IaC, and telemetry pipelines
• Clear ability to document designs for global site execution
• Willingness to travel to pod sites in US/Europe for bring‑up and changes

Responsibilities

• Own end‑to‑end network architecture for the AI Factory (scale‑up and scale‑out fabrics)
• Design optical layer, link budgets, and reliability targets for inter‑site links
• Define reference network designs so every new pod ships with validated fabric
• Build and bring up first fabrics, tune routing, congestion control, QoS, and GPUDirect RDMA
• Coordinate with inference engine, NVIDIA, and vendors to align traffic patterns
• Lead capacity planning and vendor evaluations; set telemetry and runbooks
• Hire and lead network engineers as the footprint grows to tens of thousands of ports

Skills

InfiniBand

NVLink

NVSwitch

Optical networking

DWDM

RoCE

GPUDirect RDMA

Python

Go

Infrastructure as code

Telemetry

Capacity planning

Tools

NVIDIA Spectrum-X

Quantum-X800

OpenFabrics

NCCL

Job description

Runware seeks an AI Factory Network Architect to own the design, build, and operation of the global AI supercomputer network. You will lead hands‑on fabric bring‑up, set standards, and collaborate with hardware, software, and vendor teams across US and Europe.

You will drive scale‑out and scale‑up fabrics using NVLink/NVSwitch, InfiniBand, RoCE, and optical links, while ensuring fault‑tolerance, telemetry, and incident runbooks for tens of thousands of ports.