About

Megh Thakkar is a talented Research Scientist and ML professional with 2.86 years of relevant experience in the field. Currently, he is working as a Pre Doctoral Fellow at Google, where he is developing speech-text models for a billion users. Prior to joining Google, Megh worked as a Machine Learning Engineer at Avaamo, where he worked on custom voice cloning using limited user speech data and semantic parsing and question answering over tabular data. Additionally, he has also worked as a Research Intern at MIDAS: Multimodal Digital Media Analysis Lab, where he worked on time-aware modeling of online text streams leveraging the hyperbolic space for financial analysis. He has also worked on cross-modal model agnostic interpolative data augmentation, which resulted in two research papers: HypMix: Hyperbolic Interpolative Data Augmentation (EMNLP 2021) and DMix: Adaptive Distance-aware Interpolative Mixup (ACL 2022). Megh holds a Bachelor's degree in Computer Science from Birla Institute of Technology and Science, Pilani, which he completed in 2020. He also completed his Higher Secondary education in Physics, Chemistry, and Mathematics from New Tulip International School in 2016, and his Secondary School education from Hillwoods School in 2010. Megh's technical skills include research, machine learning, and data analysis. He is a highly motivated and passionate individual who is always eager to learn and stay up-to-date with the latest developments in his field. With his expertise and experience, Megh is a valuable asset to any organization he works with.