Senior HPC Infra Engineer for Research & AI Training

PhysicsX Ltd.; Singapore; Hybrid; SGD 120,000 - 180,000; Full time

Benefits offered by this job

Hybrid work model

Shoreditch office presence

Job summary

PhysicsX Ltd. in Singapore is seeking a Senior Infrastructure Engineer to design, build, and operate our SLURM-on-Kubernetes platform supporting large-scale CAE and data-generation workloads.

You will own cluster reliability, optimize throughput across hundreds of GPUs, and collaborate with researchers to translate their needs into robust infrastructure. A hybrid work model with Shoreditch office presence is offered.

Qualifications

• 4–6 years of infrastructure/platform/HPC engineering experience.
• Hands-on with SLURM and Kubernetes in production environments.
• Experience containerizing workloads and orchestration for compute-heavy workloads.
• Strong debugging across scheduling, networking, and job failures.
• Comfortable collaborating with researchers to translate needs into improvements.

Responsibilities

• Design, build, and maintain SLURM-on-Kubernetes infrastructure for large-scale training and data generation workloads.
• Own cluster reliability, efficiency, and GPU utilization.
• Partner with Researchers to debug job failures, diagnose bottlenecks, and improve iteration speed.
• Build and maintain containerised environments (Docker, Enroot/Pyxis) for reproducible workflows.
• Manage high-performance storage systems to support data-intensive CAE workloads.
• Develop tooling, automation, and observability to reduce operational overhead.
• Contribute to capacity planning and infrastructure roadmap as the cluster scales.
• Build and implement a scaling strategy for HPC use in large-scale model training and data generation.

Skills

Infrastructure engineering

SRE / Platform engineering

Scripting / automation

Tools

Docker

Kubernetes

SLURM

Enroot/Pyxis

RDMA/NCCL networking

Job description

PhysicsX Ltd. in Singapore is seeking a Senior Infrastructure Engineer to design, build, and operate our SLURM-on-Kubernetes platform supporting large-scale CAE and data-generation workloads.

You will own cluster reliability, optimize throughput across hundreds of GPUs, and collaborate with researchers to translate their needs into robust infrastructure. A hybrid work model with Shoreditch office presence is offered.