Soham Deshmukh’s Homepage

I am 2nd year Masters student at Electrical and Computer Engineering, Carnegie Mellon University. My research interest lies in the intersection of Audio, Multi-Task Learning and Graph Signal Processing. At CMU, I have worked with Prof. Bhiksha Raj and Prof. Rita Singh on Weakly Labelled Audio data and Speech analysis. Prior to this, I received B.Tech from Department of Electrical Engineering, Veermata Jijabai Technological Institute (VJTI), where I worked with Prof. Faruk Kazi on Threat modelling and Attacker Behaviour profiling.

I would be interning at Microsoft as Data & Applied Scientist with Yammer team for summer 2020. I spent my undergrad summers interning at Siemens and Siemens R&D and competing on Kaggle where I won gold and silver medals in competitions: Google, Quora.


[Aug 2020] Last semester’s research paper and code of ‘Multi-Task Learning for Interpretable Weakly Labelled Sound Event Detection’ is now available online here

[Aug 2020] Had great summer intern experience improving and deploying Machine Learning models at scale at Yammer, Microsoft

[May 2020] I TA’ed for 18661 Introduction to Machine Learning for Engineers (Instructor: Gauri Joshi and Carlee Joe-Wong) in Spring 2020. Great course for a learning foundational ML methods (website)

[April 2020] Contribute to our COVID system led by Prof. Rita and Prof. Bhiksha by donating your voice sample here and helping us CMU researchers in improving our system. The work has appeared in multiple news, some coverage: Forbes, dailymail UK, Android Authority, Gizmodo.

[April 2020] We beat existing Weakly Supervised Sound Event Detection System by a margin (paper and code coming soon)


If I am not working on my laptop, I enjoy spending time reading fiction books and watching thriller movies. During my undergrad, I got up to speed with Shonen mangas and animes like Naruto, Bleach, Death Note, Steins;Gate etc.

Let me know if you have any good thriller fiction book or movie recommendation.