Network Solutions Architect, AI Factory Services

Lenovo seeks a Network Solutions Architect to join the Hybrid Cloud Solutions and AI Offering Engineering team within SSG. This senior-level role designs, deploys, and validates high-performance network infrastructure for Lenovo’s AI Factory and GigaFactory service offerings, covering both enterprise-scale and rack-scale GPU environments powered by NVIDIA Spectrum-X Ethernet and InfiniBand fabrics. The architect produces network reference architectures, deployment runbooks, and field-ready engineering documentation consumed directly by Lenovo Professional Services and Managed Services teams globally. This includes GPU cluster fabric design, BlueField DPU offload architectures, multi-tenant network isolation, and performance-validated RDMA and RoCEv2 configurations for AI and HPC workloads. The ideal candidate brings expert-level depth in NVIDIA Spectrum-X, Quantum InfiniBand, and BlueField DPU architectures, with proven experience designing networks for large-scale GPU clusters, high-density liquid-cooled environments, and multi-tenant AI Factory deployments. This role is based in Morrisville, NC, and includes some customer-facing responsibilities.

In this role, you will:

• Design GPU cluster network architectures for NVIDIA AI Factory environments: Spectrum-X Ethernet (Spectrum-4 SN5600 + BlueField-3 DPUs) for enterprise deployments and Q3400 XDR InfiniBand + Spectrum 5600 Ethernet for rack-scale GigaFactory deployments.
• Develop and validate RDMA/RoCEv2 configuration guides and deployment runbooks for AI and HPC workloads on NVIDIA GPU clusters.
• Architect multi-tenant network designs for NeoCloud and enterprise AI Factory customers, including namespace isolation, east-west traffic segmentation, and site resiliency aligned to NVIDIA Cloud Partner Reference Architecture.
• Design BlueField-3 DPU deployment architectures using DOCA for network function offload, security services, and service mesh in multi-tenant GPU environments.
• Define rail-optimized InfiniBand topologies and Spectrum-X Adaptive Routing, SHARP in-network computing, and congestion control configurations for AI training and inference workloads.
• Develop network architectures for high-density, liquid-cooled rack-scale GPU deployments, incorporating power-aware design patterns for performance optimization.
• Create operational runbooks for network provisioning, troubleshooting, and performance validation, consumed directly by PS and MS field teams.
• Publish network performance benchmarks, including RDMA throughput and RoCEv2 latency baselines, to support field delivery quality and pre-sales sizing.
• Collaborate with ISG product and engineering teams to validate fabric designs against NVIDIA Enterprise Reference Architectures and hardware roadmaps.
• Stay current with emerging NVIDIA networking technologies and contribute to Lenovo’s AI infrastructure innovation roadmap.