I am a Research Scientist in the GenAI Speech team at Meta, where I work on speech capabilities for LLaMA models. My research interests lie in the application of machine learning methods for speech and language tasks.

Previously, I was a PhD student at Johns Hopkins University, working in the Center for Language and Speech Processing (CLSP), advised by Sanjeev Khudanpur and Dan Povey. I was a JHU-Amazon AI2AI fellow, a Fred Jelinek fellow, and an IEEE Rising Star in Signal Processing.

When I’m not doing ML, I like to work out, climb boulders, play guitar, and read fiction.
Updates:

  • July 2024: Our AI Speech team becomes a part of GenAI, and is tasked with developing speech for LLaMA models.

  • January 2024: I joined Meta in NYC as a Research Scientist! I will be working on robust on-device ASR in the AI Speech & EMG team led by Mike Seltzer.

  • January 2024: I defended my PhD! You can find the slides and video on the Talks page.

  • November 2023: I presented a poster about the next-generation Kaldi toolkits at the NSF CIRC PI meeting in Salt Lake City.

  • September 2023: I have been awarded a Fred Jelinek fellowship by Johns Hopkins, for the academic year 2023-24.

  • June 2023: I will be spending this summer in Le Mans (France), participating in JSALT 2023. Our team will be working on WFST+end-to-end methods for speech.

  • June 2023: I was selected as an ICASSP Rising Star in Signal Processing.

  • May 2023: GSS paper accepted at InterSpeech 2023. This implementation is used in the baseline for the CHiME-7 DASR challenge.

  • February 2023: 2 papers accepted at IEEE ICASSP 2023. These papers investigate target-speaker ASR using transducers (work done at Meta AI), and using self-supervised models (led by my colleague Zili Huang).

  • October 2022: I am selected as a recipient for the inaugural JHU+Amazon AI2AI fellowship for 2022-23.

  • May 2022: I passed my GBO (JHU CS qualifying exam) and officially became a Ph.D. candidate (here are the slides for my presentation). Also, I'll be starting an internship at Meta AI (Menlo Park) in the Speech team.

  • January 2022: 2 papers accepted at IEEE ICASSP 2022. These papers investigate multi-talker ASR with neural transducers, and adding domain knowledge for fine-tuning of large self-supervised models. Here is a poster describing both papers.

  • Janurary 2022: I participated in the Mini SCALE workshop organized by HLTCOE. I was in the "Improving speech analytics for room audio" team led by Matthew Wiesner.

  • June 2021: 4 papers accepted at INTERSPEECH 2021. Check out publications page for more info! Also, I am attending ICASSP 2021 virtually :)

  • April 2021: Our JHU-GoVivace team placed 2nd (and 1st in the Hindi-English task) in the Indic code-switching challenge.

  • March 2021: I will be interning (virtually) with Dr. Jinyu Li at Microsoft this summer.

  • January 2021: Our Hitachi-JHU team obtained 2nd best DER in the Third Dihard challenge. We used several systems, and combined their outputs with a modified version of DOVER-Lap. Register for the workshop for more details!

  • November 2020: 4 papers accepted at IEEE SLT 2021. Check out publications page for more info!

  • August 2020: I will be a TA for Intro to HLT in the fall.

  • June 2020: I am participating in JSALT 2020. I will be working on informed target speaker ASR with Marc Delcroix and Shinji Watanabe.

  • May 2020: Our JHU submission to the CHiME-6 challenge obtained second-best results in Track 2 (diarization + ASR track). The system description paper is available here.

Referral policy: Please check out my referral policy before reaching out to me for referrals.