Sourya joined Galois in 2020 as a research engineer with a machine learning focus. He obtained his Ph.D. from the University of Southern California immediately prior, focusing on complexity reduction in deep learning. He first developed pre-defined sparsity to simplify neural networks, then created the AutoML framework Deep-n-Cheap to search for networks trading off performance with complexity. Prior to that, he obtained his bachelor’s degree from the Indian Institute of Technology Kharagpur in 2014.
Sourya’s research interests (which change frequently) currently include social impacts and interpretability of machine learning systems. On a lighter note, he enjoys swimming and soccer.