Technical Support Engineer – AI Cloud

Superpower X AI (Singapore) Technology Pte.; Cyberjaya; On-site; MYR 120,000 - 180,000; Full time

Job summary

Superpower X AI (Singapore) Technology Pte. seeks an experienced infrastructure support engineer to manage customer tickets end-to-end, onboard workloads on clusters, and diagnose job failures with clear root-cause analysis. The role emphasizes reliable SLA adherence and transparent customer updates.

You will monitor GPU/network/storage health, triage node faults, and collaborate with the network team on fabric issues. Proactively tune alerts, automate diagnostics, and drive incident reviews.

Qualifications

• 3+ years in infrastructure support, SRE or support engineering in a customer-facing, SLA-driven environment.
• Working knowledge of Kubernetes and/or Slurm: debug pods and jobs, read logs, follow runbooks.
• Experience with monitoring stacks such as Prometheus and Grafana.
• Scripting in Bash and Python.
• Ansible, Terraform or other infrastructure-as-code tools.
• Clear written communication: ticket notes should be understandable to engineers and customers.

Responsibilities

• Own customer tickets and channels end to end, within SLA, and keep customers informed until resolution.
• Onboard customers and run workloads on clusters: access, Slurm or Kubernetes jobs, containers, storage.
• Debug customer job failures and performance issues; explain root cause and next steps.
• Write incident updates and post-incident reports for affected customers.
• Monitor GPU, network, storage and node health across fleet; act on alerts.
• Triage GPU node faults using nvidia-smi, DCGM and system logs; isolate issues across GPU, NIC, firmware and server.
• Run InfiniBand/RoCE fabric diagnostics and work with the network team on issues.
• Drain, repair and return nodes to service; prepare evidence for vendor RMA.
• Lead or support incident response and perform root cause analysis on major incidents.
• Tune alerts for clear meaning and action; automate diagnostics and repetitive fixes.
• Feed recurring issues back to engineering.

Skills

Kubernetes/Slurm debugging

Monitoring Prometheus/Grafana

Bash/Python scripting

IaC (Ansible/Terraform)

Infrastructure/SRE experience (3+ yrs)

Clear written communication

Job description

What you'll do

Own customer tickets and channels end to end, within SLA, keeping customers informed until resolution

Help customers onboard and run workloads on our clusters: access, Slurm or Kubernetes jobs, containers, storage

Debug customer job failures and performance issues, and explain root cause and next steps clearly

Write incident updates and post-incident reports for affected customers

Cluster monitoring and stability

Monitor GPU, network, storage and node health across the fleet, and act on alerts

Triage GPU node faults using nvidia-smi, DCGM and system logs; isolate issues across GPU, NIC, firmware and server

Run InfiniBand/RoCE fabric diagnostics and work with the network team on link and fabric issues

Drain, repair and return nodes to service; prepare clean evidence for vendor RMA

Lead or support incident response, and run root cause analysis on major incidents

Operational improvement

Tune alerts so each one has a clear meaning and a clear action

Automate diagnostics and repetitive fixes with scripts and small tools

Feed recurring customer and fleet issues back to the engineering team

What we're looking for

3+ years in infrastructure support, SRE or support engineering, in a customer-facing, SLA-driven environment (cloud, data platform or managed services)

Working knowledge of Kubernetes and/or Slurm: can debug pods and jobs, read logs, follow runbooks

Experience with monitoring stacks such as Prometheus and Grafana

Scripting in Bash and Python

Ansible, Terraform or other infrastructure-as-code tools

Clear written communication: your ticket notes let the next engineer pick up without starting over, and customers understand what happened

Working arrangement

Based in Malaysia, with regular travel to our data centers across Southeast Asia

On-call rotation required