Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
ЗП не указанаразвернуть
На похожих ролях по рынку
$98к$127кмедиана$158к
23 в выборке
87
Хорошая вакансия
развернуть
Это чётко очерченная senior-роль с практической ответственностью за крупную AI- и HPC-платформу; при этом очень глубокие и узкие требования подойдут специалистам с редким профильным опытом.
Редкое сочетание навыков
Современная инфраструктура
AI-инфраструктура
Понятная зона ответственности
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Senior HPC Platform Engineer (HPC / GPU Infrastructure) (m/f/d) (HPC/GPU infrastructure): Building and operating bare-metal and virtualized GPU platforms for large-scale AI training and inference with an accent on Linux systems, NVIDIA GPUs, BlueField DPUs, InfiniBand, and storage networking. Focus on tuning kernel and PCIe/NUMA performance, operating GPU virtualization and RDMA fabrics, automating fleet lifecycle, and leading root-cause analysis across European datacentres.
Location: London, United Kingdom; flexible home-office work with an international virtual team.
Company
hirify.global is a European Neocloud operating large-scale NVIDIA GPU infrastructure for AI training and inference, with a focus on sustainable datacentres and accelerated compute.
What you will do
Build, tune, maintain, and harden Linux bare-metal GPU hosts from BIOS/BMC through kernels, drivers, firmware, NUMA, PCIe, and IOMMU.
Deploy and operate NVIDIA BlueField-2/3 DPUs, ConnectX adapters, InfiniBand fabrics, OpenSM, UFM, RDMA, and GPU cluster networking.
Deliver GPU virtualization using PCIe passthrough, SR-IOV, KVM/QEMU, libvirt, and KubeVirt on Kubernetes.
Operate IPv6-first host and storage networks supporting NVMe-oF, RoCE, NFS, block, and object traffic across spine-leaf fabrics.
Automate provisioning and lifecycle management with PXE, cloud-init, Redfish/IPMI, Ansible, Bash, and Python tooling.
Lead production incident response and root-cause analysis while coordinating hardware generations with networking, storage, Kubernetes, and customer-facing teams.
Requirements
Typically 7+ years of Linux systems or infrastructure engineering experience, including recent hands-on GPU or HPC platform work.
Deep Linux and kernel knowledge, including modules, drivers, sysctl and boot parameters, hugepages, NUMA, CPU pinning, IOMMU/VFIO, PCIe topology, dmesg, perf, ftrace, and crash analysis.
Production Debian/Ubuntu experience and comprehensive NVIDIA GPU knowledge covering drivers, CUDA, Fabric Manager, NVLink/NVSwitch, MIG, DCGM, and GPUDirect RDMA.
GPU virtualization experience with passthrough, SR-IOV Virtual Functions, vGPU concepts, and KVM/QEMU/libvirt or KubeVirt integration.
Extensive hands-on BlueField-2 and/or BlueField-3 experience is required, including BFB images, firmware, DOCA, OVS/ASAP2 offload, and host-to-DPU networking.
InfiniBand experience with both OpenSM and UFM is required, including subnet management, partitions, routing, diagnostics, RDMA concepts, perftest, and NCCL tests.
Strong IPv4/IPv6 networking fundamentals, including SLAAC/DHCPv6, neighbour discovery, routing, DHCP, DNS, NTP, and BGP in spine-leaf datacentres.
Knowledge of NVMe-oF, RoCE, NFS, multipathing, QoS, PFC/ECN, MTU, and storage traffic across multi-tier spine-leaf networks.
Excellent Bash and Python; Ansible, BMC, Redfish, IPMI, PXE, preseed, cloud-init, Docker/containerd, Kubernetes, Operators, CNIs, CSIs, and proven production incident leadership.
Clear English communication is required for collaboration across teams and datacentres in different countries.
Nice to have
VAST Data and/or Pure Storage experience.
Flatcar Container Linux or Fedora CoreOS experience.
Nautobot, OpenBao or HashiCorp Vault, and Defguard experience.
Knowledge of Slurm, MPI, NCCL, distributed training, and inference frameworks.
NVIDIA GPU Operator and Network Operator experience on Kubernetes.
Experience operating GPUs at scale for a cloud provider, Neocloud, or large HPC centre across multiple sites.
Culture & Benefits
Flexible home-office work with an international and virtual team.
Hardware support for advanced HPC work.
Idea-led environment with direct input into departments, infrastructure, processes, and culture.
Work focused on sustainable datacentres, cooling, accelerated compute, and carbon neutrality.
Wellbeing, diversity, and inclusion initiatives.
Будьте осторожны
Если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого. Это мошенники. Обязательно жмите «Пожаловаться» или пишите в поддержку. Подробнее в гайде →