ML Accelerator Architect and Performance Engineer, Silicon

Google is seeking an ML Accelerator Architect and Performance Engineer to join their Silicon team, focusing on developing custom silicon solutions for future Google products. This role combines hardware architecture and machine learning expertise to push the boundaries of on-device ML applications.

The position requires deep technical knowledge in neural network models, hardware architecture, compilers, and software stacks. You'll be responsible for driving hardware architecture exploration while collaborating with research teams and system architects to optimize future workloads. The role involves working with cutting-edge ML technologies and contributing to compiler, runtime, and API strategies.

As part of Google's hardware team, you'll be developing solutions that power millions of consumer products worldwide. The role offers the opportunity to work with talented researchers and engineers across Google, contributing to innovations in hardware-software co-design for machine learning applications.

Key responsibilities include initiating new feature modeling in architecture simulators, optimizing performance through collaboration with researchers and application developers, and enhancing user experiences by working with full-stack software engineers. The ideal candidate will have at least 5 years of relevant experience, strong programming skills in C++/Python, and deep understanding of ML frameworks like TensorFlow/JAX/PyTorch.

This position represents an opportunity to shape the future of Google's hardware experiences, delivering unparalleled performance, efficiency, and integration. You'll be at the forefront of developing custom silicon solutions that enable next-generation machine learning applications, working in a collaborative environment that pushes the boundaries of what's possible in hardware-software co-design.