Shashi Kumar

Shashi Kumar is a doctoral student in the Doctoral Program in Electrical Engineering (EDDEE) at École Polytechnique Fédérale de Lausanne (EPFL), affiliated with the School of Engineering (STI) and the IDIAP Research Institute (LIDIAP). He holds the role of Doctoral Assistant at LIDIAP, contributing to research in speech technology and machine learning. His work focuses on advancing automatic speech recognition (ASR), optimal transport frameworks, and variational autoencoders for speech enhancement and signal processing.

Research interests include speech recognition systems, multimodal task unification, far-field speech processing, and machine learning applications in signal processing and computer vision. His publications highlight contributions to SLAM-ASR performance analysis, joint speaker change detection, and PCB defect classification using image segmentation techniques.

Shashi's research is anchored at the IDIAP Research Institute, where he collaborates on projects involving deep learning, audio signal processing, and speech technology. While no awards or grants are explicitly listed, his work reflects active engagement in international challenges like the Interspeech DiCOVA competition.