Get a reply from this employer
Your resume and cover letter, tailored to exactly what they’re hiring for.
Acceler8 Talent is seeking a hands-on NPU Architect for a remote role focused on designing AI accelerators and in-memory compute architectures. You will translate workload and software requirements into architecture, microarchitecture, and implemented models, working with hardware, software, and verification engineers.
The role involves implementing features in a C++ simulation framework, analyzing compute utilization and data movement, and guiding design decisions for workloads including LLMs,
Remote | Up to $280,000 USD base
Help design a new class of AI accelerator built around precise analog in-memory computing.
We’re working with a semiconductor startup that raised an oversubscribed $100M Series B led by Tiger Global and has launched its first commercial accelerator. The team is hiring a hands-on NPU Architect to shape how future hardware executes AI workloads efficiently.
You’ll translate workload and software requirements into architecture, microarchitecture and implemented models. Your scope will include in-memory compute, on-chip networks and memory orchestration.
Working with hardware, software and verification engineers, you’ll:
Ideally, a PhD with 2-8 years of relevant experience. Strong MS researchers and engineers with substantial implemented architecture projects are also welcome.
AI accelerator, GPU and CPU architecture backgrounds are relevant. You should be able to explain what you personally built, why you chose the design and how an AI workload executes on hardware.
C++ and Python experience, digital-design fundamentals and HDL familiarity are valuable. Prior analog in-memory computing experience is not required.