2+ years in NOC, data center operations, or IT support role
Basic Linux system administration skills
Familiarity with monitoring tools like Prometheus or Grafana
Experience with ticketing systems such as ServiceNow or Jira
Ability to perform physical data center tasks like rack and stack
Strong communication skills for documentation and handoffs
Curiosity about automation and process improvement
Comfort with structured data management
Responsibilities
Monitor GPU cluster health and environmental sensors
Respond to alerts and execute runbooks for incidents
Perform hardware triage and identify failures
Execute standard remediation tasks for common issues
Collect diagnostic data for escalation to L2/SME
Manage incident tickets from creation to resolution
Perform physical data center tasks like cable installation
Maintain and update operational runbooks
Benefits
Opportunity for growth into SME roles or automation authors
Structured training environment with a focus on automation
Collaborative team culture with a focus on continuous improvement
Hands-on experience with cutting-edge AI and GPU technology
Flexible shift scheduling with a focus on work-life balance
Full Job Description
Position Overview
You are the first human in the loop - the escalation target when the AIOps system needs a decision, and the source of ground truth that turns novel incidents into new automations.
NeoCloud is building an AI-operated GPU cloud. That doesn't mean fewer humans - it means humans focus on judgment calls the platform can't yet make, and every judgment call trains the platform to do it next time. In this L1 role you cover front-line monitoring and incident response for NeoCloud's US GPU DCs during the 8AM-8PM PST shift. You execute SOPs, escalate the hard cases, and feed the AIOps substrate the ground truth it needs to learn from novel incidents.
What you'll own
Monitor GPU cluster health, network status, storage systems, and environmental sensors via centralized dashboards.
Respond to alerts and execute runbooks for common incidents: GPU errors, link flaps, node failures, storage alerts.
Perform hardware triage: identify failed GPUs, NICs, PSUs, disks, and cables from monitoring data and physical inspection.
Execute standard remediation: GPU reset, node drain/reboot, link re-seat, BMC recovery.
Collect diagnostic data for L2/SME escalation: logs, DCGM output, network diagnostics, hardware health reports.
Manage incident tickets from creation through resolution or escalation (ServiceNow/Jira).
Perform physical DC tasks: cable installation, hardware swap-outs, rack and stack, labeling (on-site roles).
Execute structured shift handoffs at 8AM and 8PM PST with the APAC operations team.
Maintain and update operational runbooks based on recurring issues.
Assist with hardware deployment, firmware updates, and inventory management under SME guidance.
Feed the AIOps substrate
Every novel incident you resolve is data the platform team needs - you tag it, describe it, and hand it back so it becomes an automation.
Every runbook you touch should get closer to being executable by the platform, not by you.
Your handoff notes are structured signal, not free-form email.
Why this role is different from a NOC job
You are not the last line of defense - the platform is. You are the training signal.
Growth path is real: strong L1s here move into SME roles, or into the platform team as automation authors.
Job Requirement:
2+ years in NOC, data center operations, or IT support role
Basic Linux system administration (command line, log analysis, service management)
Familiarity with monitoring tools (Prometheus, Grafana, Nagios, or equivalent)
Experience with ticketing systems (ServiceNow, Jira Service Management)
Ability to perform physical data center tasks: rack and stack, cabling, hardware replacement
Strong communication skills for shift handoffs, incident documentation, and escalation
Ability to work 8AM-8PM PST shift schedule (12-hour shifts with rotation)
Curiosity about automation - you don't just execute the runbook, you notice when it's the third time this month and ask what should change.
Comfort with structured data - you understand that how you file a ticket matters, because it may train a model that decides how the next one is filed.
