Job Description
Role Summary\n
Design, deploy, and validate the compute, storage, in-band, out-of-band, carrier, and customer networks supporting the GPU environment. Own joint topology and configuration decisions with Supermicro, NVIDIA, carriers, security teams, and the customer.
Core Responsibilities\n
· Produce physical and logical designs for InfiniBand and/or Spectrum-X Ethernet, storage, management, BMC, tenant, internet-edge, and private-connectivity networks.
\n
· Confirm switches, DPUs, NICs, optics, copper/fiber media, rack placement, cable lengths, patching, IP plan, routing, segmentation, redundancy, and failure domains.
\n
· Coordinate Supermicro factory and field teams on integrated rack networking and coordinate customer security, addressing, DNS, NTP, identity, firewall, and connectivity requirements.
\n
· Build controlled configurations, automated backups, inventory, telemetry, alerting, configuration validation, and change procedures.
\n
· Execute link, topology, bandwidth, latency, congestion, failover, isolation, carrier, and end-to-end workload tests; resolve defects before customer acceptance.
\n
· Deliver as-built diagrams, port maps, cable records, configuration repository, spares list, and troubleshooting runbooks.
Required Qualifications\n
· 6 or more years designing and operating data-center or HPC networks in production.
\n
· Strong experience with high-speed Ethernet and/or InfiniBand, routing, switching, optics, structured fiber, telemetry, and automation.
\n
· Familiarity with NVIDIA ConnectX, BlueField, Spectrum-X or Quantum fabrics and multi-tenant security segmentation.
\n
· Strong documentation, packet-level troubleshooting, performance testing, and vendor/customer coordination skills.