Senior AI Networking Architect — Large-Scale GPU Clusters

Hut 8 Corp.; United States; On-site; USD 180,000 - 280,000; Full time

Job summary

Highrise AI is seeking a Principal AI Network Engineer to lead end-to-end network fabric bring-up for large-scale GPU clusters and to optimize performance for AI workloads. You will own topology, routing, and automation across Ethernet and InfiniBand fabrics, coordinating with compute and storage teams for production readiness.

The role requires senior-level troubleshooting, design and deployment of high-speed networks, and close collaboration with site engineers.

Qualifications

• Bachelor's degree in computer science, engineering, networking, or related discipline, or equivalent practical experience.
• 10+ years of hands-on data center or high-performance networking experience, including deep troubleshooting and technical leadership.
• Demonstrated success deploying and operating large-scale, low-latency networks for AI, HPC, cloud, or similarly demanding distributed systems.
• Strong expertise in Ethernet and/or InfiniBand fabrics, RDMA/RoCE, routing and switching, and multi-tier data center network architectures.
• Experience with high-speed 200/400/800G networking, NICs, optical interconnects, performance testing, and network telemetry.
• Proficiency with Linux networking and network automation using Python, Ansible, or equivalent tooling; familiarity with NVIDIA networking and fabric‑management tools is highly desirable.

Responsibilities

• Lead network commissioning of materially complete AI clusters, including switch initialization, configuration validation, host connectivity, and readiness for production workloads.
• Configure and integrate high-speed Ethernet and/or InfiniBand compute fabrics, including RDMA/RoCE, alongside storage, management, and out-of-band networks.
• Validate topology, routing, redundancy, addressing, port mappings, optics, cabling, and interoperability across switches, NICs, and supporting systems.
• Implement and refine network automation, configuration management, firmware/OS baselines, and repeatable fabric deployment procedures.
• Partner with compute and storage engineers to resolve host-to-fabric integration issues and ensure end-to-end cluster readiness.
• Execute fabric qualification and acceptance testing for link health, error rates, bandwidth, latency, congestion, resilience, and scaling behavior.
• Troubleshoot and tune RDMA transport, congestion control, QoS, PFC/ECN, routing, and switch/NIC settings as appropriate to the deployed fabric.
• Analyze NCCL and distributed workload results with compute engineers to isolate communication bottlenecks and improve multi-node performance.
• Use telemetry, switch counters, fabric management platforms, and diagnostic tools to identify degraded links, optical issues, packet loss, or unexpected congestion.
• Own networking acceptance evidence, defect remediation, configuration sign-off, and readiness recommendations for commercial service.

Skills

Ethernet fabrics

InfiniBand fabrics

RDMA/RoCE

Linux networking

Network troubleshooting

Python

Ansible

NVIDIA networking tools

Education

Bachelor's degree in computer science, engineering, networking, or related discipline

10+ years hands-on data center or high-performance networking experience

Tools

Python

Ansible

NVIDIA fabric-management tools

Job description

Highrise AI is seeking a Principal AI Network Engineer to lead end-to-end network fabric bring-up for large-scale GPU clusters and to optimize performance for AI workloads. You will own topology, routing, and automation across Ethernet and InfiniBand fabrics, coordinating with compute and storage teams for production readiness.

The role requires senior-level troubleshooting, design and deployment of high-speed networks, and close collaboration with site engineers.