GPU Cloud Network SRE - InfiniBand & RoCE Expert

Doist

Austin, San Jose (TX, CA)

Hybrid

USD 140,000 - 210,000

Full time

9 days ago

Get more replies from employers

Send a job-specific resume in minutes.

Job summary

Bitdeer is seeking an experienced data center networking engineer to operate InfiniBand/Fabric fabrics and RoCEv2 networks across scalable GPU clusters in its US data centers. You will convert telemetry into actionable insights to prevent congestion and topology issues, enabling high-performance AI workloads.

You will own fabric monitoring, QoS and performance tuning, and collaborate with Nvidia/Mellanox support for escalations and firmware lifecycles.

Qualifications

5+ years in data center networking.
Hands-on experience deploying and operating Nvidia/Mellanox InfiniBand switches at scale.
Strong understanding of IB subnet management, partitioning, and QoS.
Experience with RoCEv2 deployment including PFC, ECN, DCQCN configuration.
Proficiency with UFM or equivalent IB fabric management tools.
Knowledge of 400G/800G optics, cabling standards, and structured cabling best practices.

Responsibilities

Operate InfiniBand fabrics (fat-tree/rail-optimized/dragonfly) for large GPU clusters.
Manage RoCEv2 networks across Nvidia, Arista, and Cisco devices for RDMA workloads.
Use UFM for monitoring, diagnostics and subnet management.
Tune IB/RoCE performance (routing, congestion control, traffic isolation).
Tune NCCL communication (topology detection, ring/tree algorithms, GDR).
Oversee firmware lifecycles for IB switches and HCAs.
Diagnose faults (link flaps, symbol errors, drops, routing issues).
Coordinate with Nvidia/Mellanox support for escalations and RMA.
Feed telemetry (ibdiagnet, perfquery, ibstat, PortRcvErrors, etc.) into AIOps platform.