Campus SRE: AI Compute Reliability Lead

SpaceX Super Intelligence; Memphis (TN); On-site; USD 140,000 - 180,000; Full time

Job summary

SpaceXSI is building a world-leading campus reliability team to oversee the full reliability stack across compute, network, storage, and power at one of the largest AI training campuses. You will drive incident leadership, fleet-scale observability, and blameless postmortems while coordinating with NOC and hardware teams.

Ideal candidates bring experience from power plants, data centers, or industrial facilities and demonstrate strong automation and playbook development skills to reduce MTTR and

Qualifications

• Bachelor's degree in Systems Engineering, CS, or EE
• 5+ years in SRE, systems engineering, or large-scale production ops
• Proven incident command experience on a bridge or control room
• Monitoring/observability design at fleet or campus scale
• Scripting in Python or Bash; one systems language (C/C++, Java, Go, Rust)

Responsibilities

• Own monitoring architecture across GPU/host telemetry and facility telemetry
• Provide SEV command support: incident leadership, NOC coordination, paging/escalation
• Run blameless postmortems and drive corrective actions to closure
• Gate high-risk changes: firmware/BIOS/BMC rollouts with Redfish/canaries/rollback
• Lead cross-functional reliability projects spanning compute, network, storage, and facility signals

Skills

Prometheus/Grafana

NVIDIA DCGM

SCADA/BMS/EPMS

Education

Bachelor's degree or equivalent

Tools

Redfish

IPMI

Job description

SpaceXSI is building a world-leading campus reliability team to oversee the full reliability stack across compute, network, storage, and power at one of the largest AI training campuses. You will drive incident leadership, fleet-scale observability, and blameless postmortems while coordinating with NOC and hardware teams.

Ideal candidates bring experience from power plants, data centers, or industrial facilities and demonstrate strong automation and playbook development skills to reduce MTTR and