Regional Cluster SRE — GPU/HPC Infra Lead

iFrame; Toronto; On-site; CAD 292,000 - 473,000; Full time

Job summary

iFrame seeks a hands-on Cluster Site Reliability Engineer to own the physical reality of a GPU-centric platform across seven regions. This on-site role in Toronto involves bringing up new racks, validating InfiniBand fabric, and ensuring the SLA while being the on-call owner for regional incidents.

You will work with Linux, Kubernetes, Terraform, Prometheus/Grafana, and Go or Python, coordinating with procurement and regional teams.

Qualifications

• Five-plus years operating large compute clusters in HPC or hyperscale environments.
• Deep InfiniBand and Ethernet RoCE experience with fabric debugging.
• Strong Linux fundamentals and kernel tuning knowledge.

Responsibilities

• Bring up new racks, cabling, NCCL/RCCL tests.
• Drive InfiniBand fabric to spec and reduce bit-error budget to zero.
• Run capacity planning across seven regions and coordinate with procurement.
• Own regional incident response with 4-hour resolution target.
• Build and maintain bring-up runbook for future hires.
• Manage global pager rotation and on-call duties.

Skills

Compute clusters

InfiniBand/Network tuning

Linux fundamentals

Go or Python

Calm under load

Tools

Kubernetes

Terraform

Prometheus

Grafana

Job description

iFrame seeks a hands-on Cluster Site Reliability Engineer to own the physical reality of a GPU-centric platform across seven regions. This on-site role in Toronto involves bringing up new racks, validating InfiniBand fabric, and ensuring the SLA while being the on-call owner for regional incidents.

You will work with Linux, Kubernetes, Terraform, Prometheus/Grafana, and Go or Python, coordinating with procurement and regional teams.