Achieving near-peak training throughput, particularly beneficical in memory or communication constrained environments
Hi, I'm Andrew Sheinberg. I am a third-year CS PhD student at Princeton University advised by Professor Kai Li. Prior to Princeton, I graduated from Yale where where I studied Computer Science & Mathematics. I grew up in Barrington, RI.
I am researching high-performant, machine learning systems. More specifically, I focus on understanding the relationships between operating systems and accelerator + networking software stacks. I'm developing a collection of efficent, robust, and portable software libraries to empower resource-concious machine learning workloads. In particular, this means maximizing compute utilization under memory or communication constraints. This entails investigating the low-level details of:
There is a link to some side projects I've worked on over the years. I enjoy traveling and have linked some of my favorite photos.