Senior SRE for GPU Clusters - Night Shift

HR POD Careers; Lahore; On-site; PKR 2,678,000 - 4,687,000; Full time

Job summary

HR POD Careers is seeking a seasoned SRE to own the reliability and performance of production Linux GPU clusters in Lahore. You will lead complex multi-node troubleshooting, manage GPU drivers, networking, and storage, and automate provisioning and observability.

You will work with Kubernetes/Slurm, IaC tooling, and AI-assisted automation to reduce toil and improve SLAs; prior HPC experience and open-source contributions are a plus. This role operates on a night shift from 8 PM to 4 AM.

Qualifications

• 5+ years of experience in systems, infrastructure, or SRE engineering.
• Bachelor's degree in Computer Science.
• Deep Linux troubleshooting across OS, networking, storage, and performance with root access on production systems.
• Ubuntu experience highly relevant as primary OS.
• Hands-on GPU server ops in production; driver and hardware issues.
• Practical network troubleshooting incl. physical-layer faults.
• Strong automation with Python or similar language.
• Experience with configuration management and IaC (Ansible, Terraform etc.).
• Observability/alerting with Grafana and Prometheus.
• GPU cluster / AI infra production experience.
• Kubernetes or Slurm in production; experience with both is a bonus.
• Background in HPC or research computing.
• Familiarity with NVIDIA GPU stack, InfiniBand/RDMA, NCCL.
• Experience with CLI-based AI coding agents like Claude Code.
• Open-source contributions in cloud-native, HPC, or AI infra ecosystems.

Responsibilities

• Own reliability, availability, and performance of production Linux GPU clusters (OS, drivers, GPUs, networking, storage).
• Lead end-to-end troubleshooting of complex distributed systems, GPU nodes, networking, and storage.
• Triage GPU NCCL and rail performance across multi-GPU/nodes.
• Diagnose network faults end to end including cabling and link errors.
• Configure/manage workload managers (Kubernetes, Slurm) and dependent services.
• Build automation/tooling using Python/Go to reduce toil.
• Leverage AI tools like Claude to accelerate automation and incident analysis.
• Automate provisioning, image deployment, and remediation using Ansible/IaC.
• Design observability with Grafana/Prometheus/Loki and tune meaningful alerts.
• Lead incident response, on-call, postmortems, and reliability improvements.
• Collaborate with Platform/Systems Eng for capacity planning and rollouts.

Skills

Deep Linux troubleshooting

Python

Go

Automation mindset

HPC/Research computing

NVIDIA GPU stack

Education

Bachelor's degree in Computer Science

Tools

Kubernetes

Slurm

Ansible

Terraform

Grafana

Prometheus

Loki

Claude Code

Job description

HR POD Careers is seeking a seasoned SRE to own the reliability and performance of production Linux GPU clusters in Lahore. You will lead complex multi-node troubleshooting, manage GPU drivers, networking, and storage, and automate provisioning and observability.

You will work with Kubernetes/Slurm, IaC tooling, and AI-assisted automation to reduce toil and improve SLAs; prior HPC experience and open-source contributions are a plus. This role operates on a night shift from 8 PM to 4 AM.