Aurosweta
Mahapatra

I am a Ph.D. student at Johns Hopkins University, working with Dr. Berrak Sisman on speech security. My interest in AI began during my undergraduate years, when an internship and a published paper on COVID-19 detection from radiological images showed me how impactful AI could be in real-world applications. This curiosity led me to UCLA, where I earned my M.S. and began working in speech research at the SPAPL Lab with Dr. Abeer Alwan, focusing on automatic speech recognition for child speech. Now at Johns Hopkins, my research addresses a critical gap in speech deepfake detection: existing models often struggle with emotional and expressive synthetic speech generated by advanced text-to-speech and voice conversion systems. I build emotion- and prosody-aware speech deepfake detection models that generalize across paralinguistic variation, with the broader goal of supporting the secure and ethical use of synthetic speech. I am also broadly interested in speech technologies for healthcare, speech generation, and speech translation, and I am always happy to connect with others sharing similar interests.