Senior Network Architect – AI Infrastructure

Nscale; United States; On-site; USD 120,000 - 150,000; Full time

Job summary

Nscale is looking for a Senior Network Architect located in the United States to design scalable, production-ready GPU infrastructure tailored for AI applications. This role involves collaborating with AI-native companies and hyperscalers to create robust architectures that meet complex performance and reliability requirements. The ideal candidate has over 6 years of experience in infrastructure engineering, deep knowledge of GPU systems, and exceptional communication skills. Join Nscale and help build the future of AI infrastructure.

Qualifications

• 6–10+ years in solutions architecture, infrastructure engineering, or AI/HPC environments.
• Strong knowledge of GPU-based systems and distributed training infrastructure.
• Experience with backend networking (InfiniBand, RoCE, high-speed Ethernet).
• Solid understanding of storage architectures for AI workloads.

Responsibilities

• Engage AI-native startups to translate business and ML requirements into infrastructure specifications.
• Architect end-to-end GPU cluster solutions including GPU selection, backend networking, and storage.
• Work with hyperscale partners to align on hybrid deployments.
• Partner with internal teams to ensure designs are executable.

Skills

Solutions architecture

Infrastructure engineering

Customer-facing communication skills

GPU-based systems knowledge

Distributed training infrastructure

Tools

InfiniBand

RoCE

CUDA

Job description

About Nscale

Nscale is building high-performance GPU infrastructure purpose-built for AI. We partner with AI-native companies and hyperscalers to deliver scalable, reliable, and performant compute environments for training and inference at scale. Our customers push the limits of distributed systems. We operate at the intersection of GPUs, high-speed networking, storage architecture, and production AI workloads.

The Role

We are hiring a Senior Network Architect to work directly with AI customers and hyperscale partners to translate workload requirements into scalable, production-ready infrastructure designs. You will bridge customer ambitions with Nscale’s capabilities—designing solutions across GPU compute, backend fabric, frontend networking, storage systems, and cluster architecture. This is a highly technical, customer-facing role requiring deep infrastructure knowledge and strong communication skills.

What You’ll Do

Customer Technical Discovery

• Engage AI-native startups, enterprises, and hyperscalers to understand:
  • Training vs inference workloads
  • Model size and scaling strategy
  • Distributed training topology
  • Data pipeline and storage patterns
  • Performance, latency, and reliability requirements
• Translate business and ML requirements into infrastructure specifications.

Solution Architecture & Design

• Architect end-to-end GPU cluster solutions including:
  • GPU selection and sizing
  • Backend networking (InfiniBand / RoCE / Ethernet fabrics)
  • Frontend networking and connectivity
  • Storage (parallel file systems, object storage, NVMe tiers)
  • Rack density and data center constraints
• Produce HLD/LLD documentation and reference architectures.
• Validate feasibility within Nscale’s product and operational capabilities.

Hyperscaler & Partner Collaboration

• Work with hyperscale partners to align on connectivity, interconnect, and hybrid deployments.
• Design solutions that integrate with public cloud networking and storage architectures.
• Evaluate peering, bandwidth, and redundancy requirements.

Internal Collaboration

• Partner with infrastructure engineering, deployment, and operations teams to ensure designs are executable.
• Provide feedback into product and platform roadmaps based on customer needs.
• Support pre-sales efforts with technical validation and design assurance.

Required

• 6–10+ years in solutions architecture, infrastructure engineering, or AI/HPC environments.
• Strong knowledge of GPU-based systems and distributed training infrastructure.
• Experience with backend networking (InfiniBand, RoCE, high-speed Ethernet).
• Solid understanding of storage architectures for AI workloads.
• Experience designing large-scale compute clusters.
• Customer-facing experience with strong technical communication skills.

Preferred

• Experience working with hyperscalers (AWS, Azure, GCP) or large colocation providers.
• Familiarity with NCCL, RDMA, CUDA, and distributed training frameworks.
• Experience producing formal architecture documentation.
• Understanding of cost modeling and capacity planning.

What Success Looks Like

• Customer requirements are translated into clear, scalable designs.
• Proposed architectures are technically sound and executable.
• Nscale wins complex AI infrastructure opportunities through technical credibility.
• Customers trust Nscale as a long-term infrastructure partner.

For information on how Nscale handles candidate personal data, please see our Employee & Candidate Privacy Notice: Here.