Soham Deshmukh

Applied Scientist, Microsoft


I am an Applied Scientist on the Microsoft Speech team. My research aims to build and improve the audio capabilities of machines.

Previously, I received my masters degree from Carnegie Mellon University and B.Tech from VJTI. I have been an intern at Microsoft, Siemens R&D, and Siemens. My Microsoft research page can be found here.

Research opportunities: If you have questions or want to collaborate with me, feel free to email me.


Dec 2023 Three papers accepted at ICASSP 2024 [1,2,3]
Sep 2023 🐧 Pengi is accepted at NeurIPS 2023: here
Mar 2023 Two papers accepted at ICASSP 2023: [1, 2]
Feb 2022 Scheduler service integrated in M365 and reached general availability: here
Jun 2021 Paper proposing methods for improving sound event detection in noisy environments accepted at INTERSPEECH 2021

selected publications

  1. NeurIPS 23
    Pengi 🐧: An Audio Language Model for Audio Tasks
    Soham DeshmukhBenjamin ElizaldeRita Singh, and Huaming Wang
    arXiv preprint arXiv:2305.11834 2023