Aug 2026 — present
Staff Research Scientist
Hippocratic AI
For over a decade I've worked on speech technology: keeping it trustworthy by detecting speech deepfakes and verifying who is really speaking, and making sense of conversations by tracking who spoke when. My teams have taken first place in international challenges, and my research spans industry and academia — Naver Clova, Senior Research Scientist at Apple, and now Staff Research Scientist at Hippocratic AI, with a postdoc and now an adjunct faculty appointment at Carnegie Mellon's Language Technologies Institute. Today I'm increasingly focused on building speech LLMs to be personalized and to understand paralinguistics.
Now — Staff Research Scientist at Hippocratic AI, and Adjunct Faculty at LTI, CMU.
Focus — speech LLMs, speech deepfake detection & speaker recognition.
Publications — 90+ at ICASSP & INTERSPEECH.
Community — co-organizer of the SASV, VoxSRC, ASVspoof 5 & WildSpoof challenges.
Detecting speech deepfakes and verifying who is really speaking, so voice interfaces can be trusted.
Knowing who spoke when, and making sense of multi-speaker audio.
Personalized spoken-language models that understand paralinguistics.
Full list of 90+ publications available in the CV.
I've had the privilege of working with extremely talented students, fostering their growth as researchers. These collaborations have produced impactful publications across diverse speech-processing tasks. I'm always excited to mentor new people — reach out if you'd like to explore working together.
Dynamic pruning of LLMs
Spoken language understanding · spoken dialogue systems
Robust automatic speaker verification
Speaker verification · spoofing-robust ASV
For collaboration, questions, or correspondence on speech, audio security, and machine learning:
[email protected]