Senior Hardware System Architect — Dense AI Compute

think; Riyadh; Sur place; SAR 700 000 - 1 000 000; Plein temps

Avantages offerts par ce poste

Ownership and impact

Early-stage compensation

Senior team collaboration

Résumé du poste

Think AI in Riyadh seeks a Senior Hardware Architect to own the architecture of multi-accelerator server platforms from concept to production. You will lead PCIe topology, GPU-to-NIC affinity, and cross-vendor RDMA integration, guiding the platform through EVT, DVT, and PVT while coordinating with ODMs and in-house teams.

You will drive signal integrity, power/thermal design, BIOS/BMC requirements, and platform validation.

Qualifications

• 12+ years in hardware or system architecture (GPU/accelerator focus).
• Experience shipping dense multi-GPU server platforms from architecture through production.
• Strong PCIe Gen5/Gen6 and RDMA knowledge; bottleneck analysis and topology design.

Responsabilités

• Own the system architecture end to end: CPU platform, PCIe topology, switch partitioning, GPU-to-NIC affinity.
• Design for GPUDirect RDMA and peer-to-peer traffic across heterogeneous accelerators; validate IOMMU and P2P.
• Own signal integrity strategy across retimers, PCIe channels, backplanes, and channel loss budgets.
• Lead power and thermal architecture for dense accelerator configurations, including liquid cooling readiness.
• Define BIOS/UEFI and BMC requirements; coordinate bring-up, telemetry, firmware, and fabric integration.
• Write platform specifications and reference designs; lead design reviews of schematics, layouts and BOM.
• Select and manage ODM/JDM partners for boards; drive EVT, DVT, PVT to production in Riyadh.
• Lead move from JDM to in-house board design; own schematics, layout, and SI/PI for one platform at a time.
• Own validation and quality strategy: EVT/DVT/PVT criteria, memory qualification, interop, and long-duration stress tests.
• Drive DFx with manufacturing for test coverage, burn-in, and yield targets in Riyadh.
• Lead failure analysis from power-on through mass production; produce corrective actions and guides.
• Build validation and test automation in Python/Bash linked to ILM telemetry for scalable testing.
• Build and grow hardware platform team over time in board design, SI/PCIe, mechanical/thermal, firmware, and validation.

Connaissances

Hardware architecture

GPU-accelerated servers

System bring-up

Schematic reviews

Python automation

Bash scripting

Outils

Python

Bash

Description du poste

Think AI in Riyadh seeks a Senior Hardware Architect to own the architecture of multi-accelerator server platforms from concept to production. You will lead PCIe topology, GPU-to-NIC affinity, and cross-vendor RDMA integration, guiding the platform through EVT, DVT, and PVT while coordinating with ODMs and in-house teams.

You will drive signal integrity, power/thermal design, BIOS/BMC requirements, and platform validation.