About

Yen-Jung Chang is a highly accomplished ML Research Scientist with over 7 years of relevant experience. He is currently working at Facebook, where he is focused on applied research in the areas of distributed training, inference efficiency, and model compression for super large deep models. Prior to joining Facebook, Yen-Jung was a Staff AI Software Engineer at LinkedIn, where he led the efforts on building the company-wide anomaly detection system for large-scale time series and metrics. He also developed the explainable AI and model understanding library for LinkedIn's model developers, and led the efforts on building the distributed training architecture for LinkedIn's deep learning models. Yen-Jung is an expert in automated data analytics on data cubes, and has developed an algorithm that powers one of LinkedIn's enterprise products - Strategic Talent Flow. He has also successfully detected security attacks that hide in large volumes of traffic data, helping model developers to narrow down online issues of deployed AI models. Yen-Jung holds a Ph.D. degree from the Electrical and Computer Engineering Department at the University of Texas at Austin, where his Ph.D. work focused on distributed and parallel computing, verification and debugging of distributed systems, and lattice theory. In addition to his academic achievements, Yen-Jung has won several awards, including the 1st place in the International Global Routing Contest by International Symposium on Physical Design in 2008, Best Thesis Award by Taiwan IC Design Community in 2008, Scholarship Award by SpringSoft in 2008 & 2010, NSF Student Award by National Science Foundation in 2014, and Best Teaching Award by UT-Austin in 2015. Yen-Jung is a highly skilled Research Scientist and Software Engineer, with expertise in ML, AI, Deep Learning, and Senior Software Engineering. He is an excellent problem solver with a proven track record of delivering successful projects. With his extensive experience and expertise, Yen-Jung is a valuable asset to any organization.