Research

I studied EECS at UC Berkeley and worked at the Berkeley Speech Group inside Berkeley Artificial Intelligence Research (BAIR), advised by Dr. Gopala K. Anumanchipalli, Peter Wu, Cheol Jun Cho, and Jiachen Lian. My work there covered visualizing the vocal tract with speech-driven 3D avatars, brain-computer interfaces for virtual embodiment, and generating realistic audio-visual dialogue with avatar-based LLMs.

Email, Resume, Google Scholar

• Multimodal Pretraining for Vocal Tract Modeling
  Tejas S. Prabhune, Bohan Yu, Rishi Jain, Peter Wu, Gopala K. Anumanchipalli
  Paper pending, 2024. Paper

• Towards Streaming Speech-to-Avatar Synthesis
  Tejas S. Prabhune, Peter Wu, Bohan Yu, Gopala K. Anumanchipalli
  ASRU Demo, 2023. Paper

• Contrastive learning-based pretraining improves representation and transferability of diabetic retinopathy classification models
  Minhaj Nur Alam, Rikiya Yamashita, Vignav Ramesh, Tejas S. Prabhune, Jennifer I. Lim, Robison Vernon Paul Chan, Joelle Hallak, Theodore Leng, Daniel Rubin
  Nature Scientific Reports, 2023. Paper