My research interests lie at the intersection of machine learning and high performance computing (HPC). I focus on large scale distributed training of deep neural networks, with an emphasis on both data and model parallelism. The design of machine learning software on HPC systems often presents unique challenges that can often not be addressed at research universities, and much of the cloud and HPC infrastructure used in industry is either not publicly available or is sold as a service. My work at the lab sits between these two extremes, where open source meets heavy compute power.
Last Updated: March 1, 2021 - 2:37 pm