Principal AI Back-End Network Systems Lead

Quick Facts

• Role: Principal AI Back-End Network Systems Lead

• Focus: End-to-end back-end network product leadership for AI scale-up and scale-out systems

Description

Own the back-end network product definition and execution plan across scale-up and scale-out—from requirements and architecture through development, qualification, deployment, and fleet learning. Lead cross-layer architecture and integration spanning silicon, NICs/DPUs, PHYs/retimers, switch ASICs, optics/electrical channels, and fabric topology, translating AI training/inference needs into measurable system targets.

Drive validation, reliability, manufacturing readiness, and operational learning to ensure the solution can be safely deployed and scaled in hyperscale datacenters.

Responsibilities

• Own end-to-end product ownership and architecture for back-end networking

• Translate AI training/inference requirements into system targets (bandwidth, latency, scale, reliability, availability, serviceability, power/thermal, cost, schedule, security)

• Lead NUDD technology assessments, prototypes, and risk mitigation (make/buy/partner decisions)

• Architect and integrate accelerator networking silicon, NICs/DPUs, SerDes/PHY, switch ASICs, cables/backplanes, optics, and fabric topology

• Define scalable scale-up/scale-out topologies and behaviors (radix, reach, routing, path diversity, oversubscription, failure domains, resiliency)

• Oversee optical/electrical integration (signal integrity, link training, FEC, reach, power/thermal, manufacturability)

• Define/validate protocol behavior (Ethernet/RDMA and other scale-up/scale-out protocols): flow/congestion control, QoS, routing, packet loss/recovery, telemetry, security

• Connect network architecture to distributed AI workloads and collectives (all-reduce, all-gather, reduce-scatter, all-to-all)

• Build analytical/simulation models for topology, congestion, collective efficiency, power, cost, and scaling

• Own end-to-end validation and production qualification (functionality, interoperability, performance, scale, resiliency, security, serviceability, upgrade/recovery)

• Lead bring-up and root-cause analysis across physical links to protocol/runtime/collectives/application behavior

• Define manufacturing test, deployment, monitoring, rollback, and fleet-health requirements using operational learning

• Provide technical leadership across architecture, hardware/firmware, optics, networking, software, validation, manufacturing, datacenter, and workload teams

Benefits

• Eligible for benefits and other compensation (role-dependent)

• Base pay range across the U.S.: USD $142,800–$274,800 per year (location-adjusted ranges for specific metros)