Hardware Operations Lead

Crux AI

Palo Alto (CA)

On-site

USD 180,000 - 280,000

Full time

5 days ago

Be an early applicant

Application generator

Turn this role into an interview — a resume and cover letter built around what this employer wants.

Get past ATS filters

Benefits offered by this job

Health and dental coverage

Vision coverage

Life insurance

401(k) with company match

Bonuses

Job summary

Crux AI in Palo Alto is building an integrated AI infrastructure. You will own everything inside the rack across Crux’s fleet—the TPU hosts, switches, cabling, and liquid‑cooling manifolds—to become a validated, customer‑ready cluster on the scheduled date.

The role runs the full hardware lifecycle: inbound logistics, rack integration, power‑on, break/fix, spares, and decommissioning at Crux sites and colo partners.

Qualifications

10+ years in data center hardware or server operations.

Responsibilities

You will own fleet hardware operations across Crux’s campuses and colos.

Skills

Data center operations

Hardware deployment

Team leadership

Vendor management

DCIM/CMDB

Tools

NetBox

Job description

Built to set the gold standard for integrated AI infrastructure

Crux AI is a newly formed, U.S.-based integrated AI infrastructure company created to remove the physical and operational constraints on consequential AI ambitions. Crux brings together power, high-density data centers, TPU silicon, networking, orchestration software, and ongoing operations as one integrated system.

Crux is being capitalized to plan every layer together, develop each one to demanding standards, and operate the whole system with efficiency and reliability. That gives hyperscalers, frontier AI labs, sovereign customers, enterprises, and AI-native companies greater freedom to pursue the AI they are here to create.

WHAT YOU'LL DO

You will own everything inside the rack, physically, across Crux’s entire fleet — the TPU hosts, network switches, optics, cabling, and liquid-cooling manifolds that arrive on a truck and have to become a validated, customer-ready cluster on the date Sales committed. The role runs the full hardware lifecycle: inbound logistics and receiving, rack integration and structured cabling, power‑on and validation handoff, break/fix and RMA, spares and asset records, and decommission at generation transitions — at Crux‑owned campuses and at colocation sites where you hold a partner’s remote‑hands team to our standard. You will build this function from zero: the deployment playbook, the spares program, the DCIM discipline, and the site technician teams that inherit them. Where Facility Operations owns the power and cooling that reach the rack, and the SRE & Software Engineering team owns the provisioning, health, and remediation software, you own the hands, the parts, and the records — a node the software has quarantined is a ticket in your queue until a technician has swapped the part and it has passed validation again. You will work hand‑in‑glove with Google’s hardware and support organizations, Capacity Delivery, Construction, Network Engineering, Facility Operations, the SRE & Software Engineering team, and Strategic Supply.

In this role, you will:

Build and lead the fleet hardware operations organization — site hardware leads, deployment technicians, and break/fix technicians — with staffing templates, shift and on‑call coverage models, and training and certification standards that scale ahead of the capacity ramp; hire people better than you at their craft.
Own cluster deployment end to end: inbound logistics and receiving inspection, staging and secure storage, rack integration and elevations, structured fiber and copper cabling to a published standard, liquid‑cooled rack installation and manifold connection with Facility Operations, power‑on, and validation handoff to the SRE team — treating each cluster as a gated schedule from dock to first customer workload.
Run break/fix and RMA for the fleet: repair queue and SLAs by severity, parts swaps and node re‑validation, RMA processing with Google and OEM/ODM partners, and failure‑rate trending by component and generation that feeds back into Supply and Google.
Own the spares program and asset lifecycle: spares depot sizing per site against failure rates and lead times, inventory accuracy, serialized asset tracking from receipt through decommission in DCIM (NetBox or equivalent), and reconciliation to Finance’s fixed‑asset register.
Own decommission and generation transitions: rack decommission procedures, data‑bearing device sanitization and destruction with chain‑of‑custody, asset disposition (ITAD, return, or redeploy), and the physical sequencing of TPU generation swaps inside live facilities.
Manage colocation and partner remote‑hands teams to Crux’s hardware standard — scopes of work, SLAs, access and escort procedures, and an audit program with witnessed work — and enforce remedies when performance slips.
Author and enforce the hardware operations standards library — deployment playbook, cabling and labeling standards, MOPs for energized‑rack work, ESD and lifting safety, change control on production hardware — applied consistently across owned and partner sites.
Report fleet hardware performance to the executive team monthly — deployment velocity (racks per week versus plan), time‑to‑repair, spares fill rate, RMA aging, and inventory accuracy — and drive the corrective actions those numbers demand.

SIGNALS OF SUCCESS

60 days — Fluent in the TPU rack, cabling, and liquid‑cooling requirements set from Google; deployment playbook v1 and cabling/labeling standard drafted; DCIM platform and asset‑tracking model selected with Business Systems; staffing plan and coverage model for the first site approved with hiring in motion.
6 months — First cluster deployed dock‑to‑validation under your playbook, with racks‑per‑week tracked against plan; break/fix queue, RMA process with Google, and spares depot live at the first site; every serialized asset in DCIM and reconciled to Finance; partner remote‑hands scope and SLA in force at any colo site.
1+ year — Fleet hardware operations running as a repeatable function across multiple sites with published time‑to‑repair, spares fill rate, and inventory accuracy at or better than target; component failure trending informing Supply negotiations and Google escalations; first TPU generation transition executed with decommission, sanitization, and disposition clean; site hardware leads in place and owning their sites.

EXPERIENCES, ATTRIBUTES AND MINDSET THAT INDICATE A GOOD MATCH

Experiences

10+ years in data center hardware or server operations — deployment, break/fix, and asset lifecycle for large server and network fleets — including 5+ years leading technician teams across shifts and sites in a 24/7 environment.
Direct experience deploying high‑density AI/HPC clusters: accelerator‑dense racks, high‑radix fabric cabling (InfiniBand, RoCE, or equivalent optical fabrics), and liquid‑cooled rack installation — you have run the dock‑to‑validation sequence yourself, not just approved the plan.
Built or scaled a hardware operations program through rapid growth: deployment playbooks, spares and RMA programs, DCIM/CMDB discipline with audited inventory accuracy, and vendor or colo remote‑hands management under contractual SLAs.

Attributes

Outcome driven and hands‑on. This is a builder’s seat, not a supervisory one. You will be on the dock for the first delivery, in the cold aisle for the first rack, and on the phone with Google for the first RMA — not because there is no team, but because the person accountable for the fleet’s hardware has to be close enough to the work to change it. Then you hire people better than you at their craft and stay close enough to know when something is off before it shows up in a report.
Records are part of the work. You believe a swap that is not in DCIM did not happen, that inventory accuracy is a safety and financial control rather than paperwork, and that the spares depot is sized from failure data, not from a guess.

Mindset

Posses a high tolerance for ambiguity. Hardware will arrive before the standard is finished and the customer date will not move. You write the procedure, run it, fix it, and hand it to the next site rather than waiting for one to arrive.
Demonstrate velocity through discipline, not heroics. Racks per week is produced by staging, kitting, labeled cabling, and rehearsed sequences — and a deployment saved by a weekend of overtime is a playbook that failed. You build the system that makes the next cluster boring.
AI‑agentic first. Fluent with AI agents — or committed to becoming so quickly — and you embed them as first principles in how you and your team work, defaulting to agentic workflows before adding headcount or process.

Nice to have [Preferred, not required]:

Hardware or server operations leadership at a hyperscaler (Google, AWS, Meta, Microsoft) or neocloud (CoreWeave, Lambda, Nebius, Nscale or similar) during a multi‑site deployment ramp.
Direct TPU deployment experience, or large GPU cluster deployments (NVL-class or DGX/HGX-class racks) including burn‑in and validation handoff.
DCIM/CMDB administration (NetBox or similar), ITIL/ITSM process design, and comfort with failure‑rate and inventory analytics.
Experience with data‑bearing device sanitization standards (NIST 800‑88), ITAD vendor management, and export‑control handling for accelerator hardware.
Experience managing colocation remote‑hands and smart‑hands providers to SLA from the tenant side, including audit program design.
Safety program experience for hardware work — ESD, lifting and material handling, energized‑rack MOPs, working at height.

Salary Range Information

The annual salary range for this position has been estimated based on market data and other factors. However, a salary higher or lower than this range may be appropriate for a candidate whose qualifications differ meaningfully from those listed in the job description.

About Crux