The Rack Systems TeamRack Systems owns the rack as a product: the unit of compute that Fluidstack designs, integrates, and deploys into gigawatt-scale AI data centers.Examples of key problems the team is working on:
- Architect 150kW-class liquid-cooled racks where power, thermal, weight, and interconnect budgets all bind at once and every tradeoff shows up somewhere else in the system.
- Define and hold the interfaces between rack and facility so power, liquid, and network can each evolve without forcing a redesign of the other side.
- Keep the rack roadmap ahead of accelerator vendor roadmaps, so the next generation of dense compute lands in facilities designed before that hardware existed.
- Run a small engineering team with the review rigor of a large one, catching design faults before they reach integration.
Role Scope- Own the rack as a product for 150kW-class liquid-cooled deployments: architecture, power budget, cooling, interconnect topology, structure, and serviceability, with tradeoffs made explicitly and defended on the record.
- Write and hold the interface control documents between rack and facility, precise enough that power, liquid, and network teams can each move independently without breaking the boundary.
- Lead the rack EE, ME, thermal, and integration engineers, owning the architecture they execute and the technical calls that unblock them.
- Run design reviews where a 7-person team catches what a 70-person team would, so faults die in review rather than in the field.
- Align the rack roadmap with accelerator vendors and Fluidstack's facility architecture, so each rack generation lands in a facility built to receive it.
What We're Looking ForThe below is a starting point. We always make space for exceptional people, so if you don't fit this role exactly, tell us where you would.- 10+ years in system architecture of rack-scale compute at a hyperscaler, NVIDIA-ecosystem ODM, or AI systems company.
- You've owned a rack as a product end to end: power budget, cooling architecture, interconnect topology, weight, transport, and serviceability.
- You've shipped dense liquid-cooled rack-scale products, NVL-class or equivalent.
- You've led the EE, ME, and thermal engineers who executed your architecture, not just advised them.
- You've traded real things away on a shipping rack and taken the pushback that came with it.
- You write interface control documents that let both sides of a boundary move independently, and you've seen what happens when they don't.
- You hold a defensible view on copper versus optics for scale-up interconnect at 200G+ per lane, and you can argue where the crossover sits.
- Bonus: NVL-class rack programs, transport and seismic qualification, facility-side liquid cooling fluency, or direct engagement with accelerator vendor roadmaps.
We are committed to pay equity and transparency.
