Senior Cluster SRE: GPU Racks, InfiniBand & On‑Site
Posted Updated
Description
iFrame in Canada is seeking a Cluster Site Reliability Engineer to own the physical reality of the platform across seven regions. You will bring up new GPU racks, validate InfiniBand fabric end-to-end, and keep the cluster running at the SLA. This is a hands-on role and you will see the hardware.
Location is on-site with multiple region options. The role covers capacity planning, incident response, and creating a robust bring-up runbook for future hires, while collaborating across regions and
Bring up new GPU racks: cabling, ToR, validation, MFU tests. Drive InfiniBand fabric to spec and reduce errors to zero. Forecast demand and model power/thermal headroom across regions. Own regional incident response with defined SLA targets. Build/run the bring-up runbook for future hires. Manage the global pager and coordinate regional duties. Five-plus years operating large compute clusters. Deep InfiniBand and Ethernet RoCE experience with fabric tuning. Strong Linux fundamentals: kernel params, NUMA topology, cgroups. Comfort writing Go or Python for tooling. Calm under load; named on a $50M-ARR account when issues arise.Skills
Sign in to read and join the discussion.