Presenter

Evangelos Georganas is a Senior Research Scientist at Intel's Parallel Computing Lab in Santa Clara, CA, USA. His research interests lie in high-performance computing for deep learning and scientific applications. He has been working on tensor programming models that strive for high-performance, productivity, and portability across architectures. These benefits have been showcased within multiple contemporary deep learning workloads including Recommender Systems, Language Models, Recurrent Neural Networks, Convolution Networks, as well as within Intel’s oneDNN Deep Neural Network Library. He earned his Ph.D. in Computer Science from UC Berkeley in 2016, where he was advised by Prof. Katherine Yelick.
During his Ph.D. he worked on scalable parallel algorithms for genome and meta genome analysis, and on communication-avoiding algorithms. His work boosted the de novo assembly of the human genome, bringing the time down to 4 minutes on a supercomputer, and was recognized in 2015 by HPCwire as the best use of HPC application in Life Sciences. The high-performance algorithms he developed were used for the first whole-genome assembly of wheat, whereas his work on distributed-memory metagenome assembly was a Best Paper Finalist in SC18, and it produced the largest, high-quality metagenome assembly to date.