Hyperscale AI Network Architect

SambaNovaSystems; San Jose (CA); On-site; USD 250,000 - 330,000; Full time

Job summary

SambaNova Systems is seeking a Network Architect to lead the design of network fabrics powering hyperscale AI clusters, from RDU-accelerated servers to global deployments. You will own fabric, interconnect, and spine-leaf decisions that shape our compute platforms and ensure scalable deployments.

You will partner with hardware engineering, supply chain, and customer-facing teams to translate workload requirements into platform roadmaps, adopting a vendor-agnostic approach to avoid lock-in and

Qualifications

• 12+ years designing hyperscale, AI, HPC data center infra.
• Deep knowledge of AI cluster networking and high-speed interconnects.
• Experience with lossless Ethernet designs including QoS, ECN, PFC.
• Strong spine-leaf and underlay/overlay design expertise.
• Familiarity with high-speed optics, 400G/800G transceivers, cabling.
• Ability to translate workload requirements into product roadmaps.

Responsibilities

• Architect network fabrics powering hyperscale AI clusters end to end.
• Own fabric, interconnect, and spine-leaf design decisions for platforms.
• Collaborate with hardware engineering, supply chain, and customer teams.
• Drive vendor-agnostic selections to avoid lock-in and accelerate deployment.

Tools

InfiniBand

RoCEv2

Ethernet fabrics

400G/800G interconnects

VXLAN

EVPN

BGP

ECMP

Network automation

Job description

SambaNova Systems is seeking a Network Architect to lead the design of network fabrics powering hyperscale AI clusters, from RDU-accelerated servers to global deployments. You will own fabric, interconnect, and spine-leaf decisions that shape our compute platforms and ensure scalable deployments.

You will partner with hardware engineering, supply chain, and customer-facing teams to translate workload requirements into platform roadmaps, adopting a vendor-agnostic approach to avoid lock-in and