We are looking for a Linux Systems Administrator to manage the compute infrastructure that supports our AI research and engineering. Your core focus will be bare-metal servers running Ubuntu Linux, NVIDIA GPU systems for high-performance computing (HPC), and helping orchestrate and optimise virtual compute in the cloud.
This is a hands-on infrastructure role combining physical server administration with automation and performance troubleshooting. You will help researchers and engineers run workloads reliably, use shared resources efficiently, and choose appropriate on-premises or cloud capacity for their needs.
You will work alongside our existing IT Administrator and Ops Team, who own end-user device security. This role focuses on server and compute infrastructure, with shared responsibility for secure operations and ISO 27001 evidence.
What You Will Be Doing
- Install, configure, patch and harden Ubuntu Linux on bare-metal servers, planning maintenance to minimise disruption.
- Manage physical server health, firmware, disks, storage and remote management interfaces; diagnose hardware faults and coordinate repairs.
- Configure and maintain NVIDIA GPU drivers, CUDA compatibility and GPU-enabled container runtimes.
- Monitor GPU utilisation, memory, temperatures and errors, and troubleshoot performance across compute, storage and networking.
- Help manage shared HPC resources, including job scheduling, queues, resource limits and fair access for research and engineering workloads.
- Provision and maintain cloud virtual machines and GPU instances, using repeatable configuration and automation.
- Help orchestrate workload provisioning, scheduling and teardown across on-premises and cloud environments, taking account of data security and transfer costs.
- Optimise cloud compute through right-sizing, utilisation monitoring, automated shutdown of idle resources and cost tracking; use interruptible capacity where workloads can tolerate it.
- Maintain server networking, secure remote access, user permissions, backups and tested recovery procedures.
- Automate routine administration using Bash, Python and configuration-management or infrastructure-as-code tools.
- Maintain monitoring, runbooks, asset records and security evidence, and support incident response and capacity planning.
The Stack
Testing & CI: Molecule · ansible-lint/yamllint · pre-commit · GitHub Actions · Prometheus/Grafana
Ways of working: Git for everything · idempotent, check-mode-first changes · every change proves it works
Requirements
- Solid hands-on experience administering Ubuntu Linux servers from the command line, including services, packages, permissions, storage and logs.
- Experience managing physical servers and diagnosing hardware, operating-system and resource issues.
- Practical experience supporting NVIDIA GPUs on Linux, including driver installation and troubleshooting GPU availability or CUDA compatibility.
- Understanding of shared compute workloads and how CPU, GPU, memory, storage and network constraints affect performance.
- Working knowledge of networking fundamentals, including TCP/IP, DNS, routing, firewalls, SSH and VPNs.
- Experience provisioning and administering virtual machines in a cloud environment such as AWS, Azure or GCP.
- Ability to automate routine tasks using Bash or Python, and maintain scripts and configuration in version control.
- A disciplined approach to access management, patching, backups, change control and documentation.
- Clear communication and the ability to prioritise infrastructure issues with researchers, engineers and colleagues in IT.
Nice to Have
- Experience with HPC schedulers such as Slurm, or GPU workload orchestration using Kubernetes.
- Familiarity with Docker, Apptainer or NVIDIA Container Toolkit.
- Experience with Terraform, Ansible, cloud-init or similar provisioning and configuration tools.
- Knowledge of multi-GPU or multi-node workloads, NCCL, InfiniBand or high-speed Ethernet.
- Experience managing shared or high-performance storage and investigating I/O bottlenecks.
- Familiarity with NVIDIA DCGM, Prometheus, Grafana or equivalent monitoring tools.
- Experience with cloud cost optimisation, quotas, capacity planning or interruptible compute.
- Exposure to ISO 27001, Cyber Essentials or infrastructure operations in regulated environments.
- Ownership of the compute infrastructure supporting real-world AI research and engineering.
- Collaboration with our IT, research and engineering teams, with opportunities to deepen your HPC and cloud expertise.
- Private healthcare, including Bupa and mental health support.
- Flexible working hours and hybrid working around infrastructure requirements.
- Private pension with company contribution.
- Cycle-to-work scheme and regular team socials.
- 34 days annual leave, comprising 25 days, public holidays and your birthday.
- A culture that values autonomy, curiosity and getting things done.
#J-18808-Ljbffr
