Overview
As a Network Infrastructure Engineer, you will design and specify the physical network for AI and HPC clusters, covering out-of-band, frontend/storage, and accelerator fabrics. Youβll own from specification to installation, shaping topology, cabling, and materials for scalable, high-performance deployments. You collaborate with installers, suppliers, and project managers to ensure smooth re-cabling and reuse with minimal downtime. This role offers visibility in a growing tech company focused on cutting-edge AI research platforms. You will be pivotal in delivering robust, scalable interconnects that enable advanced workloads.
Responsibilities
- Design physical networks for new AI/HPC clusters across multiple fabrics (out-of-band, frontend/storage, accelerator) including topology, switch placement, port maps, cable schedules, and BOMs
- Specify fibre and copper infrastructure in detail (fibre grade, connectors, polarity, patch-panel layouts, loss budgets) and select matching transceivers and DACs/AOCs
- Plan changes to existing clusters to allow re-cabling and reuse with minimal disruption
- Communicate design clearly to installers, suppliers, project managers, and engineers
- Collaborate with the team on high-density, high-performance interconnects and documentation
- Evaluate and incorporate 400G/800G Ethernet and OSFP/QSFP-DD optics, optical circuit switching, and novel interconnects as relevant
- Utilise documentation tools (e.g., NetBox) and consider colocation facilities in planning
Key requirements
- Experience designing networks for HPC or AI cluster deployments, with ownership from specification to installation
- Solid understanding of out-of-band, frontend, and accelerator (GPU) fabric networking
- Practical knowledge of InfiniBand, RoCE or similar lossless fabrics and topologies (fat-tree, rail-optimised)
- Deep knowledge of data centre fibre optics (OS2 vs OM4/OM5, MTP/MPO vs LC, APC vs UPC, polarity, connectors, losses from panels/cassettes)
- Ability to clearly articulate a design to diverse stakeholders
- Experience with 400G/800G Ethernet and OSFP/QSFP-DD optics; familiarity with optical circuit switching or similar interconnects; comfortable with high-density racks and related cooling
- clear communication
- cross-functional collaboration
- problem solving
- InfiniBand
- RoCE
- lossless fabrics
Network Infrastructure Engineer β AI Platform employer: Intellectual Capital Resources
Join a dynamic team where innovation meets collaboration, as we empower our Opto Mechanical Engineers to transform cutting-edge optical concepts into tangible products. Our supportive work culture fosters continuous learning and professional growth, with hands-on opportunities in a state-of-the-art laboratory environment. Located in a vibrant area, we offer competitive benefits and a chance to work alongside industry experts, making this an ideal place for those seeking meaningful and rewarding employment.
Contact Details:
Intellectual Capital Resources Recruitment Team