Senior AI/HPC Hardware Architect – CTO Office Leader

Molex; Austin (TX); On-site; USD 140,000 - 225,000; Full time

Job summary

Molex in Austin, TX seeks a Senior Hardware Architect to drive next-generation HPC/AI hardware designs. You will own system-level architecture across electrical, thermal, and mechanical domains, and guide reference designs in collaboration with engineers, silicon partners, and product teams.

In this role you will lead SerDes, interconnect strategies, and power/thermal optimization for accelerators like GPU/TPU/IPU, shaping architectures that balance performance and reliability in data-center

Qualifications

• 10+ years of hardware design and architecture experience.
• Deep expertise in power delivery, signal integrity, and thermal management.
• Experience with AI accelerators (GPU, TPU, IPU) and multi-die systems.
• Strong collaboration and communication across multidisciplinary teams.

Responsibilities

• Define system-level requirements across electrical, thermal, and mechanical domains.
• Design scalable, reliable hardware architectures and reference designs.
• Collaborate with electrical, thermal, and mechanical teams for integration.
• Architect AI/HPC systems targeting accelerator platforms.
• Lead design reviews and promote best practices across teams.
• Stay current with emerging technologies to boost system performance.

Skills

Hardware design & architecture

Power delivery

Signal integrity

Thermal management

SerDes tech

AI accelerators

Education

Electrical/Computer Engineering degree

Tools

HFSS

Job description

Molex in Austin, TX seeks a Senior Hardware Architect to drive next-generation HPC/AI hardware designs. You will own system-level architecture across electrical, thermal, and mechanical domains, and guide reference designs in collaboration with engineers, silicon partners, and product teams.

In this role you will lead SerDes, interconnect strategies, and power/thermal optimization for accelerators like GPU/TPU/IPU, shaping architectures that balance performance and reliability in data-center