Hardware System Diagnostic Engineer

MatX develops vertically integrated silicon, hardware, and software for large-scale machine-learning workloads. The Hardware System Diagnostic Engineer owns end-to-end diagnostics for accelerator boards and racks, supporting manufacturing, debugging, and deployment of MatX’s rack-scale AI data center product.

Responsibilities:

• Own the end-to-end diagnostics test suite for accelerator boards and racks
• Develop diagnostic tools covering the AI chip, PCIe, HBM, power, and thermal
• Define test coverage and pass/fail criteria for manufacturing screening
• Enable JDM/ODM partners with documentation, tooling, and hands-on debug support
• Support partners technically through NPI and production ramp phases
• Triage and root-cause manufacturing test escapes and field failures
• Build and maintain CI/CD infrastructure for the diagnostics software stack
• Track quality metrics — yield, test coverage, defect density — and report out

Requirements:

• 7+ years in systems diagnostics, hardware validation, or manufacturing test
• Expert in Python and C/C++ for test automation and low-level device programming
• Deep knowledge of PCIe, memory hierarchies, power management, and I/O
• Proven experience managing test relationships with ODM/JDM or CM partners
• Experienced building and scaling test automation frameworks from the ground up
• Proficient with Linux system tools for hardware interrogation and fault analysis
• Self-directed and effective in a fast-moving, low-overhead startup environment
• Strong communicator across engineering, manufacturing, and executive audiences
• Experience with AI accelerator, GPU, or custom ASIC diagnostics
• Familiarity with HBM or advanced packaging (2.5D/3D) test considerations
• Exposure to rack-scale management: BMC, PDU, IPMI, or Redfish
• Experience writing diagnostics that run at bare-metal or pre-OS environments
• Familiarity with error injection frameworks and fault coverage analysis
• Prior NPI experience at a semiconductor or early-stage hardware startup
• Familiarity with Linux drivers, PCIe enumeration, or register-level programming