Чётко описанная руководящая роль в быстрорастущей AI-инфраструктурной компании даёт много ответственности и опыт работы с крупными GPU-кластерами, но требует постоянного присутствия на месте и дежурств 24/7.
Дежурства 24/7
Зона ответственности
Инфраструктура для ИИ
Щедрые льготы
Мэтч & Сопровод
Для мэтча с этой вакансией нужен Plus
Подключить Plus
Описание вакансии
Текст:
Hirify AI/Оригинал
TL;DR
Data Center Hardware Operations Site Lead (GPU): Deploying, operating, and decommissioning high-density GPU clusters across the full bare-metal server lifecycle with an accent on physical infrastructure, data center systems, and site leadership. Focus on planning rack layouts and capacity, troubleshooting hardware and network failures, coordinating contractors and facilities teams, and responding to power, cooling, and safety issues.
Location: United States; Santa Clara or Westlake. Fully on-site role with an on-call rotation supporting 24/7 operations.
Salary: $150K–$295K, plus equity.
Company
Baseten provides applied AI infrastructure, inference platforms, and developer tooling for companies deploying machine learning models in production.
What you will do
Lead rack-and-stack, cabling, turn-up, upgrades, break/fix, and decommissioning for GPU clusters.
Plan data hall white space, cabinet positions, fiber trays, rack elevations, power, cooling, and network capacity with DC Facilities.
Support data hall bring-up through site walks, drawing reviews, construction-to-operations turnover, and commissioning activities.
Diagnose hardware, optics, transceivers, link errors, power, cooling, and leak alarms; coordinate RMAs and remediation with the relevant teams.
Write and execute MOPs and SOPs, inspect installed equipment, and maintain safety and site standards.
Schedule FTE and contractor labor, coordinate cabling vendors, and manage logistics, spares, inventory, and DCIM asset data.
Requirements
Bachelor's degree in a technical field or equivalent practical experience.
6+ years of data center operations experience, including at least 2 years in a lead or supervisory role.
Hands-on experience racking, stacking, and cabling servers and network equipment at scale using copper and fiber.
Experience across the full bare-metal server lifecycle: installation, break/fix, upgrades, and decommissioning.
Comfort with Linux command-line hardware diagnostics, including tools such as ipmitool, dmesg, and lspci.
Working knowledge of data center networking, including optics, transceivers, and link-error troubleshooting.
Working knowledge of mechanical and electrical systems, including UPS, PDUs, RPPs, busway, CRAH/CRAC units, containment, liquid cooling, and BMS/EPMS alarms.
Track record of leading data center projects end to end and directing contractors and vendors across facilities, network, and engineering teams.
Sound judgment and rapid prioritization under pressure, including knowing when to escalate.
Must work on-site, lift up to 50 lbs, work in elevated locations, and participate in a 24/7 on-call rotation.
Nice to have
Experience deploying and maintaining large NVIDIA GPU clusters, including HGX/DGX platforms such as H100, H200, or B200.
Direct liquid-cooling experience with CDUs, rear-door heat exchangers, and leak response.
Construction and commissioning experience for new data halls.
Experience with InfiniBand, RoCE, Cisco IOS, Juniper Junos, or similar networking systems.
Experience with DCIM tools such as NetBox.
Python or Bash scripting for diagnostics or inventory automation.
CompTIA Server+, Network+, or similar certifications.
Experience standing up a new site from scratch.
Culture & Benefits
Competitive compensation with meaningful equity.
Flexible PTO and paid parental leave.
U.S. employees and dependents receive 100% coverage for medical, dental, and vision insurance.
U.S. employees have access to a company-facilitated 401(k).
Fertility and family-building stipend through Carrot.
Exposure to a variety of machine learning startups and production AI infrastructure.
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 →