Senior AI Systems Administrator

Reference: BH-70238Senior AI Systems Administrator | Cambridge | Permanent or Contract Most infrastructure roles stop at the software layer. This one starts at the rack. A mission-led, not-for-profit AI organisation in Cambridge needs a Senior AI Systems Administrator. Startups, enterprises, the public sector and academia share their compute and engineering effort to build AI responsibly, and you’ll be the person who keeps that compute running. You’ll provision and look after multi-rack GPU clusters used for both training and inference, working with everything from the hardware and fabric up to the orchestration layer.What you’ll be doingDeploying, tuning and managing modern GPU servers across several racksBuilding and running high-performance networks (InfiniBand, RDMA, RoCE), plus the switches, routers and firewalls around themRunning Linux at an advanced level across several distributions (Ubuntu, RHEL), including hosts running hypervisors and VMsManaging the estate as code with Git and Terraform/OpenTofuWorking closely with researchers and engineers to keep improving how the systems performWhat you’ll bringSignificant experience as a Systems Administrator, HPC Engineer, MLOps Engineer or similarHands-on GPU server deployment and tuningStrong InfiniBand/RDMA/RoCE knowledge and solid network design skillsAdvanced Linux, virtualisation, Git and infrastructure-as-code experienceA collaborative approach: you like working with the people who use the systemsBonus points for: parallel or high-availability storage (Lustre, Ceph, Weka, PEAK:AIO), Slurm or LSF, and Python scripting.Why it’s worth a lookAccess to the latest AI tooling, and encouragement to use it in how you workA growing organisation where one person’s work makes a visible differenceContacts across industry, government and academiaA competitive package with pension, benefits and professional developmentA city-centre office a few minutes’ walk from Cambridge stationSupport is UK business hours only. Major maintenance occasionally falls on a weekend, scheduled in advance, with no on-call rota. Permanent or contract considered.Senior AI Systems Administrator | Cambridge | Permanent or Contract Most infrastructure roles stop at the software layer. This one starts at the rack. A mission-led, not-for-profit AI organisation in Cambridge needs a Senior AI Systems Administrator. Startups, enterprises, the public sector and academia share their compute and engineering effort to build AI responsibly, and you’ll be the person who keeps that compute running. You’ll provision and look after multi-rack GPU clusters used for both training and inference, working with everything from the hardware and fabric up to the orchestration layer.What you’ll be doingDeploying, tuning and managing modern GPU servers across several racksBuilding and running high-performance networks (InfiniBand, RDMA, RoCE), plus the switches, routers and firewalls around themRunning Linux at an advanced level across several distributions (Ubuntu, RHEL), including hosts running hypervisors and VMsManaging the estate as code with Git and Terraform/OpenTofuWorking closely with researchers and engineers to keep improving how the systems performWhat you’ll bringSignificant experience as a Systems Administrator, HPC Engineer, MLOps Engineer or similarHands-on GPU server deployment and tuningStrong InfiniBand/RDMA/RoCE knowledge and solid network design skillsAdvanced Linux, virtualisation, Git and infrastructure-as-code experienceA collaborative approach: you like working with the people who use the systemsBonus points for: parallel or high-availability storage (Lustre, Ceph, Weka, PEAK:AIO), Slurm or LSF, and Python scripting.Why it’s worth a lookAccess to the latest AI tooling, and encouragement to use it in how you workA growing organisation where one person’s work makes a visible differenceContacts across industry, government and academiaA competitive package with pension, benefits and professional developmentA city-centre office a few minutes’ walk from Cambridge stationSupport is UK business hours only. Major maintenance occasionally falls on a weekend, scheduled in advance, with no on-call rota. Permanent or contract considered.