Вакансия из Hirify Global, списка международных tech-компаний Для мэтча и отклика нужен Plus
Пожаловаться
Зарплата и рынок
развернуть
$144к-273к/год
$180к$200кмедиана$230к
По похожим ролям (49) ~ $200к/год
На СНГ рынке: $3.3к/мес ($2.5к - $4.3к)
86
Хорошая вакансия
развернуть
Задачи чётко сосредоточены на развёртывании GPU- и HPC-инфраструктуры, описание подробное, а компания предлагает обучение; при этом работа с клиентами может означать смену проектов.
?Клиентские проекты
Актуальные технологии
Чёткие обязанности
Обучение и сертификаты
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
AI Infrastructure Engineer (NVIDIA/Linux): Building and integrating compute, networking, storage, and cluster-management layers for AI infrastructure with an accent on GPU systems, high-performance fabrics, and repeatable bare-metal provisioning. Focus on automating deployments, validating RDMA and storage performance, controlling configuration drift, and troubleshooting production data-center infrastructure.
Location: Dallas, United States of America; partially remote
Salary: $143,640–$273,000 annually for the U.S.; regional ranges may vary. Eligible for a discretionary annual bonus and comprehensive benefits.
Company
hirify.global designs, builds, manages, and modernizes mission-critical technology systems for Fortune 100 clients.
What you will do
Build and integrate compute, GPU, network-fabric, storage, and out-of-band connectivity layers according to approved designs.
Configure BIOS/BMC, firmware, Linux or DGX OS, GPU drivers, NIC/HCA/DPU components, time services, DNS, IPAM, management access, and hardware-management integrations.
Install and administer NVIDIA Base Command Manager or comparable bare-metal provisioning and cluster-management tooling.
Validate RDMA, GPUDirect RDMA and Storage, MTU, QoS, PFC/ECN, routing, link health, fabric telemetry, and high-performance storage.
Automate builds, diagnostics, remediation, and evidence capture with Ansible, Python, Bash, APIs, Git, and CI/CD; maintain golden configurations, as-built records, known issues, support bundles, and handover artifacts.
Requirements
5+ years administering Linux, servers, networking, storage, or clustered infrastructure in production data centers.
Hands-on troubleshooting across firmware, operating-system, network, storage, and hardware-management layers.
Automation experience with Ansible, Python, Bash, APIs, Git, or equivalent tooling.
Nice to have
Experience with NVIDIA DGX/HGX, BCM, DGX OS, InfiniBand, Spectrum-X, UFM, BlueField, ConnectX, or DOCA.
Knowledge of GPUDirect RDMA/Storage, NVLink/NVSwitch, NVIDIA diagnostics, and support workflows.
Experience with HPC clusters using Slurm, Bright Cluster Manager, Warewulf, xCAT, or similar tooling.
Experience with RoCEv2, high-performance Ethernet, RDMA fabrics, or large cloud GPU estates.
Experience with parallel storage and enterprise platforms such as Spectrum Scale, Lustre, BeeGFS, Weka, VAST, DDN, Pure, Dell, HPE, or NetApp.
NVIDIA networking, AI infrastructure, systems administration, Linux, vendor network, or storage certifications.
Culture & Benefits
Full-time employment with a partially remote work arrangement.
Medical and dental coverage, disability and retirement benefits, paid leave, and paid time off.
Access to learning programs and industry certifications, including Microsoft, Google, Amazon, and Skillsoft.
Inclusive workplace supported by employee inclusion networks and volunteering programs.
If working at a client, customer, or partner location, proof of vaccination may be required under that location's policy; medical or religious accommodations may be requested.
Hiring process
Applications are accepted on a rolling basis.
Applicants using an employee referral should select “Employee Referral” and provide the referring contact's hirify.global email address.
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 →