GPU Infrastructure Architect – NVIDIA / HPC

Intellectual Capital Resources Ltd; United Kingdom; Remote; GBP 90,000 - 130,000; Full time

Job summary

IC Resources is seeking specialist GPU Infrastructure Architects and Engineers with hands-on experience in designing and deploying large-scale NVIDIA compute environments, including multi-node GPU clusters and bare-metal infra.

Ideal candidates will have expertise with HGX/DGX, Hopper/Blackwell, InfiniBand RDMA, high-speed Ethernet, NVLink/NVSwitch, and GPUDirect, with a proven track record in cluster commissioning, benchmarking and cross-layer troubleshooting.

Qualifications

• Hands-on experience in designing and deploying large NVIDIA GPU infrastructures.
• Direct involvement in architecture, deployment or commissioning of multi-node GPU clusters.
• Proven knowledge of high-speed networking and bare-metal infrastructure.
• Experience with NVIDIA interconnects and AI fabrics is highly desirable.
• Ability to troubleshoot across hardware, networking and system layers.
• Familiarity with NVLink/NVSwitch and related GPU technologies.

Responsibilities

• Architect and deploy multi-node NVIDIA GPU clusters in scalable environments.
• Integrate compute, storage and networking for high-performance AI workloads.
• Commission, validate and benchmark GPU infrastructure end-to-end.
• Troubleshoot issues across hardware, network and software layers.
• Collaborate on topology design and performance optimization.

Skills

GPU infrastructure

NVIDIA HGX/DGX

Multi-node design

InfiniBand RDMA

AI fabrics

NVLink/NVSwitch

Bare-metal infra

Cluster commissioning

Performance benchmarking

Troubleshooting hardware

Tools

ConnectX

BlueField

Spectrum-X

NCCL

Job description

IC Resources is looking to speak with specialist GPU Infrastructure Architects and Engineers with experience designing and bringing up large-scale NVIDIA compute environments.

The requirement sits at the infrastructure layer of large-scale AI and HPC platforms, covering the integration of GPU compute, high-speed networking, storage and bare-metal infrastructure.

We're particularly interested in engineers who have been directly involved in the architecture, deployment or commissioning of multi-node NVIDIA GPU clusters rather than solely operating workloads once the platform is live.

Key experience:

• Large-scale NVIDIA GPU infrastructure
• NVIDIA HGX and/or DGX platforms
• Hopper and/or Blackwell architectures
• Multi-node GPU cluster design and deployment
• InfiniBand / RDMA
• High-speed Ethernet / AI fabrics
• NVIDIA networking and interconnect technologies
• Bare-metal compute infrastructure
• GPU cluster topology
• Compute, network and storage integration
• Cluster commissioning and validation
• Performance testing and benchmarking
• Troubleshooting across hardware, networking and infrastructure layers
• Additional experience with technologies such as NVLink, NVSwitch, ConnectX, BlueField, Spectrum-X, NCCL or GPUDirect would be particularly relevant.

This is a fully remote opportunity and we're open to speaking with specialist engineers internationally.