Discipline
Networking and interconnect recruiters.
Network engineers who design the fabric, fiber, and interconnect that turn racks of GPUs into a working cluster.
How this market works
What this search really turns on.
AI fabric is its own discipline
Designing a low-latency, non-blocking fabric for GPU training is not enterprise networking. We screen on whether someone has really built for collective communication at scale, not just data center networking broadly.
Optics and fiber are half the job
At cluster scale, fiber allocation, optics selection, and cabling plans are a real part of the design and a real source of cost and risk. We match on candidates who have owned that, not just the logical topology.
Vendor fluency matters
The choices across switch vendors and NVIDIA networking carry real tradeoffs in performance, cost, and lead time. We look for people who can reason about them, not just deploy one stack.
These roles sit between teams
Network design here touches compute, facilities, and procurement at once. We screen for people who can actually coordinate across those seams.
Plain answers
Questions, answered straight.
What levels do you recruit for?
Network engineers and architects focused on data center and AI cluster fabric, up through the leads who own interconnect design across a deployment.
Is this different from traditional network engineering?
Yes. The scale, the latency requirements, and the collective-communication patterns of AI training make this a distinct skill set, and we recruit for it specifically.
Designing the fabric, or ready for a bigger build?
Tell us what you are building. The first conversation costs nothing and stays between us.