AI HPC & GPU Infrastructure Architect

THE SUPREME HR ADVISORY PTE. LTD.; Singapore; On-site; SGD 56,000 - 78,000; Full time

Job summary

THE SUPREME HR ADVISORY PTE. LTD. seeks an experienced infrastructure engineer to design, deploy and manage high-density GPU compute clusters for AI workloads in Singapore.

You will optimize container orchestration (Kubernetes, Slurm, Ray) and ensure low-latency networking and scalable storage for distributed training. Candidates have 3–6+ years in HPC/DevOps, strong Linux admin skills, CUDA/NVLink familiarity, and hands-on IaC with Terraform and Ansible.

Qualifications

• Linux systems administration with kernel tuning and Bash/Python scripting.
• Deep GPU compute knowledge, CUDA runtimes, PCIe/NVLink topology.
• Hands-on with Kubernetes and HPC schedulers (Slurm, Run:ai, Ray).
• Experience with RDMA, InfiniBand, and ECN configurations.
• Familiarity with high-IOPS storage and data pipelines.
• Bachelor's degree and 3-6+ years in HPC/DevOps/infra.
• Relevant certifications are a plus (CKA/CKAD, cloud Architect).

Responsibilities

• Architect, configure, and maintain high-density GPU compute clusters.
• Implement and manage container orchestration platforms optimized for AI workloads.
• Monitor GPU health, telemetry, and thermals; minimize idle compute time.
• Design low-latency networks and scalable storage for AI datasets.
• Build automated deployment pipelines with Terraform, Ansible, Helm, or Pulumi.
• Lead incident response, RCA, and disaster recovery planning for AI environments.

Skills

Linux administration

Shell scripting

GPU architectures

Kubernetes

Slurm

Ray

NCCL latency

RDMA InfiniBand

Terraform

Ansible

Education

Bachelor's Degree in CS/IT/CE

Tools

Kubernetes GPU operator

Slurm

Run:ai

Ray

Terraform

Ansible

NVIDIA CUDA

Job description

THE SUPREME HR ADVISORY PTE. LTD. seeks an experienced infrastructure engineer to design, deploy and manage high-density GPU compute clusters for AI workloads in Singapore.

You will optimize container orchestration (Kubernetes, Slurm, Ray) and ensure low-latency networking and scalable storage for distributed training. Candidates have 3–6+ years in HPC/DevOps, strong Linux admin skills, CUDA/NVLink familiarity, and hands-on IaC with Terraform and Ansible.