HPC Systems & Scheduler Administrator

Select Source International; Greenville (SC); Hybrid; USD 90,000 - 140,000; Full time

Job summary

Select Source International in Greenville is seeking an HPC Systems Administrator to own Linux, Slurm, and VDI in GE Vernova's enterprise HPC stack. You will work within the ADT organization with the DT team to drive reliability, throughput and user experience on the scheduler and OS layer.

The role demands senior Linux administration, expert tuning of partitions and QoS, plus hands-on management of environment modules, MPI toolchains, and monitoring across Ganglia, Grafana/Prometheus, and

Qualifications

• Experience administering Linux HPC clusters and Slurm workload manager.
• Strong knowledge of Linux OS, patching, hardening and security.
• Familiarity with HPC application environments, MPI, compilers.
• Experience with Linux VDI management and user access control.
• Ability to monitor systems with Ganglia, Grafana/Prometheus, CloudWatch.

Responsibilities

• Administer Linux OS on compute/login/head nodes.
• Manage Slurm partitions, policies, QoS, reservations.
• Support environment modules, toolchains, MPI libraries.
• Maintain Linux VDI images and user sessions.
• Implement system monitoring and alerting; respond on-call.
• Tune scheduler for throughput and reduced queue times.
• Profile and optimize performance with network and I/O.

Skills

Linux Administration

Slurm Scheduler

HPC Environment

System Monitoring

Performance Tuning

Security Compliance

Tools

Ganglia

Grafana

Prometheus

CloudWatch

Environment Modules

MPI Libraries

Job description

Greenville, SC (Hybrid)

Duration: 1 Year with possible extension (W2)

Job Title

HPC Systems Administrator

Job Description

The HPC Systems & Scheduler Administrator is a senior-level technical specialist responsible for the Linux systems administration, Slurm workload management, and Linux Virtual Desktop (VDI) environment supporting GE Vernova's enterprise HPC. Embedded within the ADT organization and working in close collaboration with the DT team, the engineer will be a primary technical owner of the scheduler and Linux OS layer of the HPC stack, driving system reliability, throughput, and user experience.

Key Responsibilities

• Linux Systems Administration
  Manage OS-level configuration, patching, and hardening of HPC compute nodes, login nodes, and head nodes; ensure stability and security compliance in coordination with DT's cybersecurity policies.
• Slurm-Based Scheduler Administration
  Serve as the primary subject matter expert for the Slurm workload manager; manage partition configurations, node scheduling policies, fairshare/priority weighting, reservation management, and Quality of Service (QoS) tiers.
• HPC Application Environment
  Manage environment module frameworks (e.g., Lmod/Environment Modules), compiler toolchains, MPI libraries, and application runtime environments in support of engineering simulation codes.
• Linux VDI Management
  Administer and maintain Linux Virtual Desktop environment; manage image lifecycle, user session management, OS configuration standardization, and application delivery to Linux desktop users.
• System Monitoring & Alerting
  Deploy and maintain comprehensive system monitoring (e.g., Ganglia, Grafana/Prometheus, CloudWatch) for compute, scheduler, and Linux VDI environments; establish alerting thresholds and on-call response protocols with DT.
• Scheduler Optimization & Throughput
  Design and tune scheduling configurations to maximize throughput and minimize queue latency; develop strategy to prioritize HPC usage across user community and reduce queue times.
• Latency & Performance Optimization
  Profile and tune system-level performance bottlenecks including network fabric (e.g., EFA/InfiniBand), I/O subsystems, and memory bandwidth; collaborate with the HPC Performance Engineer on application-level profiling.
• User Account & Access Management
  Support Linux-side user provisioning, group management, SSH key governance, and sudo/privilege management in collaboration with DT's identity management processes.