Currently a research scientist at Interactions Corporation (which acquired AT&T's Speech & Language Research in December 2014)
Visiting Researcher at the Spoken Language Processing Group, Columbia University, working with Professor Julia Hirschberg
Teaching a seminar on Spoken Dialogue Systems in Spring 2015
Current work focuses on handling speech recognition errors in spoken dialogue systems by generating targeted clarification questions
Developed a localized error detection approach using prosodic, syntactic, semantic, and ASR confidence features with machine learning to identify misrecognized words
Previously worked on the CODA project with Paul Piwek at the Open University (UK) to automatically convert monologues into dialogues
Built a parallel corpus of monologues and dialogues and derived automatic conversion rules