Technical Program Manager – Networking NPI

Nscale•Seattle, WA

1d•$190,000 - $260,000

About The Position

Nscale is hiring a Technical Program Manager to lead New Product Introduction (NPI) for networking across its US GPU cloud. The role owns the process from selection of a new network platform, architecture, or vendor through qualification, documentation, and readiness for deployment at scale. This includes new switch and NIC generations, 800G and next-generation optics, cluster fabric designs, and data center and WAN reference architectures. The role works closely with Network Architecture and Engineering, Hardware Engineering, Supply Chain, vendors, and Deployment, and involves some travel to vendor labs and US pilot sites. The Technical Program Manager leads networking product qualification and validation, vendor engagement, pilot builds, and the handoff of qualified platforms to Deployment. The role also establishes NPI processes, phase gates, templates, dashboards, and reporting to help new platforms launch predictably as Nscale grows. The position is important to ensure new network platforms are proven before they are deployed at GPU cluster sites.

Requirements

5+ years in technical program management, hardware NPI, or network engineering, with direct ownership of bringing new network hardware or architectures into production.
Experience at a hyperscaler, major network or silicon vendor, or ODM/OEM, or in a comparable large-scale environment.
Hands-on familiarity with hardware qualification, including lab validation, interoperability testing, firmware and NOS lifecycle, defect triage, and phase-gate NPI processes.
Strong technical grounding in data center and cluster networking, including Ethernet and InfiniBand fabrics, RoCE, BGP, EVPN/VXLAN, optics and transceivers, and structured cabling.
Proven ability to drive vendors and cross-functional engineering teams to decisions while designs, schedules, and supply are changing.
Comfort working in ambiguity, defining processes where none exist, and making clear readiness decisions.
Excellent written and verbal communication skills, including presenting technical trade-offs to senior leadership.
Willingness to travel to vendor labs and pilot sites as needed.

Nice To Haves

NPI experience in networking for large GPU clusters, HPC, or AI training environments.
Experience qualifying 400G/800G/1.6T optics, co-packaged or linear-drive optics, or new SerDes generations.
Familiarity with HPC topologies such as Fat Tree, Rail-optimized, and Dragonfly, and NVIDIA Spectrum-X or Quantum InfiniBand platforms.
Exposure to network automation and test tooling, including Python, Ansible, and traffic generators such as IXIA/Keysight or Spirent.
Background in supply chain or manufacturing ramp for new hardware.
PMP, PgMP, or similar certification.

Responsibilities

Own the NPI program for new networking platforms, from vendor selection and early samples through lab qualification, pilot build, and release to Deployment.
Build NPI plans with phase gates, entry and exit criteria, owners, and dates.
Drive qualification and validation of switches, NICs/DPUs, optics, cables, and network operating systems, including interoperability, performance, burn-in, and failure testing.
Manage vendor engagement on new products, including roadmap alignment, sample allocation, bug and RMA tracking, firmware and NOS release readiness, and escalation of blocking defects.
Partner with Network Architecture to turn new designs into production-ready reference builds, golden configurations, bills of materials, cabling and rack standards, and acceptance test plans.
Lead first-of-a-kind pilot builds, capture lessons learned, and feed them back into designs before scale-out.
Own the handoff to Deployment, including documentation, training, readiness reviews, and a supply plan that supports the volume ramp.
Track technical, schedule, and supply risks across new product lines; make trade-offs visible and drive decisions with engineering and leadership.
Build the NPI process, including templates, gate reviews, dashboards, and reporting.