HPC Support Engineer

Overview

In this role you will support and enhance the institution’s high-performance computing environment, including HPC and GPU clusters, research storage, high-speed networking and hybrid cloud infrastructure. You will administer Linux-based clusters, manage scientific software stacks and help researchers maximise HPC resources. You will automate deployment and monitoring using Python, Bash, Ansible, Docker, Kubernetes and CI/CD tools, while collaborating with researchers and technical teams. This is your chance to contribute to cutting-edge computational research and AI-driven projects at a prestigious, research-led university in London.

Responsibilities

• Administer and support HPC clusters, GPU systems and specialist research hardware
• Manage workload scheduling with Slurm or OpenPBS
• Maintain scientific software, compilers and libraries using Spack or EasyBuild
• Automate the lifecycle of compute and GPU nodes
• Troubleshoot complex Linux, storage, networking and performance issues
• Support containerised and cloud-based research workloads
• Help researchers optimise applications for parallel computing and GPU acceleration
• Deliver technical guidance, documentation and training

Key requirements

• Advanced Linux cluster administration experience
• Experience with HPC schedulers such as Slurm or OpenPBS
• Knowledge of CUDA or OpenCL
• Strong scripting and automation skills using Python, Bash or Ansible
• Experience with Docker, Kubernetes, Git and CI/CD practices
• Knowledge of monitoring and observability tools such as Grafana, ELK Stack or Splunk
• Experience using HPC software build frameworks such as Spack or EasyBuild
• Knowledge of scientific software stacks, compilers and libraries
• Experience managing large-scale storage and file systems
• The ability to work directly with researchers and translate technical requirements into practical solutions
• Experience with cloud-based HPC, infrastructure-as-code tools, research computing environments or higher education would be advantageous

• collaboration
• clear communication with researchers
• problem-solving
• Advanced Linux cluster administration
• HPC schedulers: Slurm or OpenPBS
• CUDA or OpenCL

…