AI Infrastructure Architect GPU Compute & Networking

Role Overview

We are looking for an AI Infrastructure Architect to design and deliver GPU infrastructure, with a focus on server hardware, high-speed networking, and physical connectivity.

This is a hands-on technical individual contributor role covering infrastructure design, hardware selection, solution validation, and deployment support. You will work with equipment vendors, network engineers, and data center delivery teams to translate capacity and performance requirements into practical infrastructure solutions.

Key Responsibilities

• Design GPU infrastructure solutions covering compute servers, cluster networking, storage connectivity, and rack-level deployment.
• Evaluate and select GPU servers and supporting components, including CPUs, memory, NICs, and storage, balancing performance, compatibility, availability, and cost.
• Develop cluster network topology and connectivity plans, including switch sizing, port allocation, uplinks, and optical transceiver or cable selection.
• Produce architecture proposals and bills of materials (BOMs), working with vendors to validate component compatibility and solution feasibility.
• Provide technical guidance during installation, bring-up, and acceptance testing, and help resolve hardware or connectivity issues.
• Coordinate with data center facilities teams on rack space, power, and cooling requirements.

Essential Requirements

• At least 5 years of relevant experience in server infrastructure, HPC/AI infrastructure, data center networking, or infrastructure solution architecture.
• Practical experience designing or deploying GPU infrastructure, with clear personal responsibility for hardware selection, network design, or technical validation.
• Solid understanding of GPU server architecture and component compatibility, including PCIe, GPU interconnects, and network interfaces.
• Hands-on experience with at least one high-speed cluster networking technology: InfiniBand or RoCE, including switch connectivity and topology planning.
• Practical knowledge of optical transceivers, fiber, DAC, and AOC selection, including compatibility, distance, and deployment considerations.
• Ability to develop technical proposals and BOMs, evaluate vendor solutions, and support deployment and troubleshooting.
• Strong communication, vendor coordination, and technical documentation skills.

Preferred Experience

• Experience with NVIDIA HGX or comparable multi-GPU platforms, including NVLink/NVSwitch-based systems.
• Experience with multi-rack GPU clusters or AI/HPC deployments at scale.
• Exposure to high-density racks, liquid cooling, or infrastructure performance testing.
• Bachelor's degree in Computer Science, Engineering, Communications, or a related discipline, or equivalent practical experience.