Senior NPU Architect – Remote (US & Canada)

Job overview

The Senior NPU Architect role is described as hybrid remote, with the listing title specifying remote work in the US and Canada; the listing location is San Francisco, CA. The posting states $180,000–$240,000 per year. The role is for a client of BEARCloud and supports development of next-generation AI acceleration technology. The architect defines NPU architecture and microarchitecture, models and analyzes performance, and evaluates AI/ML workloads to guide design decisions. Collaboration across hardware, software, design, verification, and AI teams supports high-performance NPU solutions.

What you'll do

• Design NPU architecture and microarchitecture, including compute engines, memory subsystems, and on-chip networks
• develop architectural features based on AI workloads and performance needs
• build and enhance C++ simulation models
• analyze performance and optimize compute, memory movement, and hardware utilization
• evaluate LLM, diffusion, and CNN workloads
• partner on validation and testing strategies
• collaborate on hardware/software co-design
• research emerging accelerator technologies.

What we're looking for

Skills & competencies

senior npu architect

remote us & canada

hybrid remote

san francisco ca

bearcloud

ai accelerators

npu architecture

microarchitecture

c++

python

verilog

llms

$180k–$240k/year

Work arrangement

• Weekend coverage required

Why this role

The description seeks a BS or MS in a related field with approximately five to eight years of relevant experience; a Ph.D. with two to four years is a plus. Desired experience includes AI accelerators, NPUs, GPUs, SoCs, C/C++, Python, Verilog or SystemVerilog, memory architecture, and performance modeling.

About the employer

BEARCloud, Inc. is hiring for this role. Industry: Computer Facilities Management Services. Sector: 54.