AI Storage ASIC Architect for 3D XPU Accelerators

Sandisk; Milpitas (CA); On-site; USD 180,000 - 260,000; Full time

Job summary

Sandisk is seeking an AI/ML ASIC Architect to develop AI Storage Solutions based architecture and accelerator specifications. You will drive frontend architecture for AI/ML xPU-based accelerators and prototype 3D package integration with AI Storage Solutions.

You will collaborate with design and product planning, author architecture specs, and work with RTL/DV/Emulation teams to assess performance, power, and area while delivering innovative accelerators for the data-centric era.

Qualifications

• 15+ years hands-on architecture experience in ASIC/SoC design.
• Experience authoring architecture specifications for AI accelerators.
• Strong background in PCIe/UCIe/CXL and DMA engines.

Responsibilities

• Drive AI/ML ASIC architecture integrating AI Storage with accelerators.
• Write clear architecture specifications for xPU-based accelerators.
• Define I/O subsystem and DMA architectures across subsystems.
• Collaborate with design, RTL/DV, and FW teams to evaluate PPA tradeoffs.
• Guide pre-silicon design, verification, and post-silicon validation.
• Improve architecture performance via HW/SW co-optimization.

Skills

AI/ML ASIC architecture

PCIe/UCIe/CXL

DMA architectures

SoC subsystem design

HW/SW co-optimization

Heterogeneous computing

ARM/x86/RISC-V integration

Frontend architecture

VLSI/ASIC design

Education

PhD/MS in Computer Engineering or Electrical Engineering

Tools

CUDA programming

NVMe/NAND experience

UCIe/CXL world

Job description

Sandisk is seeking an AI/ML ASIC Architect to develop AI Storage Solutions based architecture and accelerator specifications. You will drive frontend architecture for AI/ML xPU-based accelerators and prototype 3D package integration with AI Storage Solutions.

You will collaborate with design and product planning, author architecture specs, and work with RTL/DV/Emulation teams to assess performance, power, and area while delivering innovative accelerators for the data-centric era.