Head of Data Center Architecture

Overview

You will own the physical architecture of Verda's data centers end to end. You will translate incoming GPU platforms (HGX systems, NVL racks, liquid‑cooled Blackwell and Rubin trays) into buildable, repeatable site designs that can be deployed quickly and operated reliably at scale.

• Define the reference design for a Verda hall: rack layout and floor plan, power distribution and busway topology, structured cabling and fiber plant, Infini Band and Ethernet fabric routing, and the airflow or liquid‑cooling loop that keeps it all within spec.
• Produce the drawings, elevations, and specifications the deployment and site teams build against.
• Own the Bill of Materials for each deployment (servers, switches, PDUs, cabling, transceivers, cable management, and every ancillary part), keeping it accurate, versioned, and costed.
• Work with procurement and our hardware partners (including Supermicro) to ensure delivered items match specifications and that lead times and substitutions never block a deployment.
• Set standards that allow us to repeat a design across sites: naming conventions, cabling schemes, labeling, rack‑and‑stack runbooks, and acceptance criteria for handover to operations.
• Drive design reviews for new sites and expansions, resolving trade‑offs between density, power budget, cooling capacity, cost, and time‑to‑live.
• Build and lead the architecture function, working shoulder to shoulder with Electrical, Thermal, Structural, Hardware, and Networking engineers, and with Data Center Operations and Expansion teams to turn designs into live capacity.
• Author the reference design for a new liquid‑cooled Blackwell/Rubin hall and take it from floor plan and BOM to first racks online; standardize the structured cabling and fiber plant across sites; run the design review and operational handover for a new‑market expansion, from lease constraints and power envelope to acceptance testing.

• Deep, hands‑on experience designing high‑density data center white space (rack layout, power distribution, structured cabling, and cooling) at scale.
• Track record deploying GPU or HPC compute (HGX / NVL / DGX‑class systems, Infini Band or high‑radix Ethernet fabrics), ideally including liquid‑cooled racks.
• Fluency with BOM ownership: building, versioning, and costing detailed hardware BOMs and working with procurement and OEM/ODM partners on lead times and logistics.
• Strong command of the drawings and documents that define a build: floor plans, rack elevations, power and cooling schematics, cable schedules, and cabling standards.
• Understanding of power and cooling fundamentals (PDU/busway topology, redundancy models, PUE, airflow management, direct‑to‑chip and rear‑door cooling).
• Experience writing the standards and runbooks that make a design repeatable across multiple sites, and running design reviews and operational handovers.
• Leadership experience building or managing a technical team, with the range to stay hands‑on in the details.
• Bias toward speed and ownership, comfortable making decisions with incomplete information and driving them to a deployed result.