# HPC Network Engineer **Company**: Fuse Energy **Location**: London **Work arrangement**: onsite **Experience**: senior **Job type**: full-time **Category**: IT **Industry**: Energy **Apply**: https://jobs.workable.com/view/dTK5ymsmbfzXo7z6j8VMRJ/hpc-network-engineer-in-london-at-fuse-energy?utm_source=yubhub.co&utm_medium=jobs_feed&utm_campaign=apply **Canonical**: https://yubhub.co/jobs/job_82fa4338-68e ## Description Fuse Energy is a forward-thinking renewable energy startup on a mission to deliver a terawatt of renewable energy - fast. The Opportunity You'll design, deploy, and operate the network fabric for our multi-tenant AI cluster. This covers the full stack: the high-performance compute and storage fabrics carrying RDMA traffic between GPUs, the tenant-facing and management networks, fire walling and tenant isolation, and the out-of-band infrastructure that keeps it all recoverable. Responsibilities - Design and operate lossless, RDMA-capable fabrics (e.g. RoCEv2, InfiniBand) for GPU compute and storage traffic, including QoS, congestion control, and buffer tuning at scale - Build and manage leaf-spine data centre fabrics, with routed underlay and overlay design (e.g. BGP, EVPN/VXLAN) - Implement and maintain per-tenant network isolation across compute, storage, and management planes - Automate network provisioning, configuration, and validation, treating switch config as code (e.g. Ansible, Python, NetBox as source of truth), deployed through CI - Build telemetry and observability for the fabric: flow-level and buffer-level visibility, dashboards, and alerting that catches congestion and link degradation before tenants do (e.g. Prometheus/Grafana/Datadog, streaming telemetry) - Troubleshoot performance issues end to end, from optics and cabling through switch buffers to NIC/DPU configuration and collective-communication behaviour on the hosts - Operate the out-of-band management network, console access, and remote recovery paths - Support tenant onboarding: segmentation and addressing, bandwidth and isolation guarantees, and capacity planning as the cluster scales - Write clear design documentation capturing decisions, rationale, and rejected alternatives - Own and maintain the office network: wired and wireless infrastructure, firewalling, VPN/remote access, and connectivity between the office and data centre environments - Upskill colleagues on networking: share knowledge through documentation, run-throughs, and pairing so the wider team can operate and troubleshoot the fabric confidently Requirements - 5+ years as a network engineer operating production data centre networks - Strong dynamic routing experience (BGP in particular), plus overlay/encapsulation design and troubleshooting (e.g. EVPN/VXLAN) - Hands-on experience with leaf-spine / Clos fabric design and operation - Experience with modern data centre network operating systems and comfortable in the Linux networking stack, not just a vendor CLI - Practical RDMA fabric experience: lossless Ethernet (e.g. RoCEv2 with PFC/ECN/DCQCN tuning) or InfiniBand, with an understanding of why lossless behaviour matters for GPU workloads - Network automation as a working practice, not an aspiration: scripting (e.g. Python), configuration management (e.g. Ansible), config generation from a source of truth, version-controlled changes - Solid Linux administration fundamentals: you can debug from the host side as well as the switch side - Experience with network telemetry and monitoring (e.g. Prometheus/Grafana, sFlow/IPFIX, streaming telemetry) - Experience running corporate/campus networks: wired and wireless, switching, NAC/802.1X, VPN and remote access (e.g. Cisco Catalyst/Meraki or comparable) - Clear communicator who enjoys teaching: able to document, pair, and run sessions that bring less network-savvy colleagues up to speed Nice to have - Experience with GPU cluster networking specifically (e.g. NVIDIA Spectrum-X or Quantum InfiniBand, ConnectX/BlueField NICs and DPUs, UFM, SHARP, or equivalent Broadcom/Arista AI fabric platforms) - Container networking experience: CNI plugins and BGP integration between clusters and the fabric - Understanding of collective-communication libraries and how fabric behaviour shows up as training/inference performance - Multi-tenant network design: VRF-based isolation, tenant bandwidth guarantees, secure shared infrastructure - Experience with enterprise firewall platforms (e.g. FortiGate, Palo Alto), including HA deployment and virtualised/segmented instances - Storage networking experience (e.g. NVMe-oF, lossless storage fabrics, per-tenant storage isolation) - Bare-metal provisioning environments (e.g. MAAS, PXE, Redfish) and how network bootstrap fits into node lifecycle - Optical layer knowledge at 200/400/800G: transceivers, MPO cabling, link qualification - Experience standing up a data centre network from greenfield - Relevant certifications (e.g. CCNP/CCIE or equivalent), valued as evidence of depth, not a gate Benefits - Competitive salary and an equity sign-on bonus - Biannual bonus scheme - Fully expensed tech to match your needs - Breakfast and dinner allowance for office-based employees ## Skills ### Required - Network engineering - Data centre networks - Dynamic routing - Overlay/encapsulation design - Leaf-spine / Clos fabric design - RDMA fabric experience - Network automation - Linux administration - Network telemetry and monitoring - Corporate/campus networks ### Nice to have - GPU cluster networking - Container networking - Collective-communication libraries - Multi-tenant network design - Enterprise firewall platforms - Storage networking - Bare-metal provisioning environments - Optical layer knowledge --- Source: [Apply at jobs.workable.com](https://jobs.workable.com/view/dTK5ymsmbfzXo7z6j8VMRJ/hpc-network-engineer-in-london-at-fuse-energy?utm_source=yubhub.co&utm_medium=jobs_feed&utm_campaign=apply)