Postdoctoral Associate, Speech & Audio Machine Learning
Department of Electrical and Computer Engineering, University of Rochester
I build efficient and trustworthy speech and audio models: neural audio codecs, voice conversion, audio watermarking, speech recognition and speaker modeling.
I am a Postdoctoral Associate at the University of Rochester, working with Prof. Zhiyao Duan on codec-based speaker attribute disentanglement, voice conversion and audio watermarking. I received my PhD in Computer Science and Computational Linguistics from Indiana University Bloomington in 2025, where I was part of Prof. Minje Kim's Signals and AI Group in Engineering (SAIGE).
My research focuses on data efficiency and model complexity reduction for speech and audio systems, using discrete representations and neural codecs. I have worked on speech recognition, speaker diarization, speech enhancement, speech synthesis and foundation models for multi-channel audio, both in academia and in industry at Rev.com, Amazon, Google and Coqui.
Earlier, as a computational linguist working with Prof. Francis Tyers, I developed low-resource speech recognition systems and finite-state morphological analysers.
Thesis: Data efficiency and model complexity reduction for speech processing systems
Committee: Minje Kim (Chair, UIUC), Francis Tyers (Chair, Indiana University), David Crandall (Indiana University), Damir Cavar (Indiana University)