Security Systems Engineer

Description

The company builds and runs data centers for AI compute, aiming to deliver 10 to 100s of GWs faster than anyone else. It acquires power, designs and builds sites, and operates them with teams spanning hardware and software.

The security team protects the model weights trained on the infrastructure, the most valuable and most targeted assets in tech, and is building the compute to hold them at a pace no one has matched. A breach would mean the frontier leaving the building.

The team is building the whole security program from zero across continents, covering physical and logical security from the hardware up to leadership.

Work is self-directed: you own problems end to end, take on scope without waiting to be asked, and act beyond your core role without asking first.

The company rejects vendor-managed, one-appliance-per-site methods common in physical security and builds versioned, reproducible, recoverable infrastructure instead.

Responsibilities

• Design the security control plane end to end, making the architecture decisions the rest of the platform will depend on before the first campus relies on it.
• Build and run that control plane at each site, on premises where resilience and bandwidth require it, with a thin cloud management layer where that works better.
• Own PKI, certificate lifecycle, identity integration, and automated deprovisioning so access ends the moment employment ends, not at the next audit.
• Build the integrations, APIs, and monitoring that show a site is degrading before anyone raises a ticket, across an estate adding campuses faster than manual processes can track.
• Replace hand-built, appliance-per-site, vendor-managed patterns with infrastructure that is versioned, reproducible, and recoverable.
• Own backup, disaster recovery, and the non-production environment where changes are proven before reaching a live site.

Requirements

• Expect a starting point, not a strict list: if you don't match this role exactly, the company wants to hear where you would fit.
• You have designed and run a production virtualization and clustered storage platform, and have walked through a failover you tested yourself rather than assumed.
• You have been the design authority for an architecture others built on, and can name a decision you got wrong and what unwinding it cost.
• You have run database high availability and recovery, and have restored from your own backups rather than trusting they ran.
• You run PKI and certificate lifecycle as automation, so expiry is a solved problem that never pages you at 2am.
• You have integrated identity systems so provisioning and deprovisioning need no human involvement.
• You have replaced a vendor-managed or hand-built system with something reproducible, and can explain what the accepted approach gets wrong and why.
• You write code and scripts as a routine part of the work.
• Bonus: video management and access control platforms (Genetec Security Center, Milestone, Lenel).
• Bonus: enterprise virtualization and clustered storage (Hyper-V, VMware vSphere, Nutanix, failover clustering).
• Bonus: database high availability (SQL Server Always On, clustering).
• Bonus: cloud management planes (AWS).
• Bonus: infrastructure as code (Terraform, Ansible, PowerShell DSC).

Conditions

Long hours at high intensity are part of the work.

The company is committed to pay equity and transparency.

Applications are screened with AI, and the screening tool undergoes a bias audit.