Simon J. Hollis

Research Scientist

I am a Research Scientist in the On-device AI team in the AR/VR group, focusing on compilation of high-level machine learning models to AI accelerator architectures. I also contribute to the open-source Glow compiler and the PyTorch machine learning framework.
I have a PhD, MA, and BA in Computer Science from Cambridge University. Between 2007 and 2015, I was an Associate Professor in Computer Science at University of Bristol, UK, specializing in multi- and many-core processing systems and energy-efficient hardware/software co-design. In 2013, I was a Visiting Professor at Carnegie Mellon University. In both my work at IBM Research, Almaden Labs, and Xilinx Inc., I architected state-of-the-art, highly parallel AI processors. I have 35+ peer-reviewed publications.


Hardware/software co-design for energy-efficient computation, compiler design and optimization, AI, many-core systems, networks-on-Chip (NoC), embedded system benchmarks