Atharva Anand Joshi

Hi everyone! I am Atharva Anand Joshi, incoming Machine Learning Engineer 2 at Hewlett-Packard - Poly. Here, I explore advanced Deep Learning-based techniques for Spatial and Personalized Speech Enhancement directly on the headset digital signal processor. Our work allows headsets to adapt to the voices of multiple users and over time get better at isolating only their speech in real-time. I am fascinated by the application of machine learning in diverse domains like audio, finance, and marketing.

I completed my M.S in Electrical and Computer Engineering with concentration in AI/ML Systems from Carnegie Mellon University. My research spans speech processing, deep learning, and representation learning, and I’ve had the privilege of collaborating with WAVLab under the guidance of Professor Shinji Watanabe. I earned my B.E in Electrical and Electronics Engineering from BITS Pilani in July, 2022. During my undergrad, I worked on research projects under the guidance of Professor Syed Mohammad Zafaruddin (SM Research Group) and Professor Ananthakrishna Chintanpalli.

Before joining CMU, I worked as an Analyst at American Express, AI Labs . Here, I explored modelling approaches involving a blend of Tabular deep learning with Tree-based algorithms for Credit Default Prediction - What is the probability of a given customer failing to repay their outstanding debt in near future? This helps in setting the credit line and making other such credit-related decisions.

During my internship at AmEx, I developed a template-based framework that allows users to seamlessly create and deploy their end-to-end Self Learning pipelines for Sequence Models. I have also interned at Adobe Research, India with mentors Dr Atanu Sinha and Dr Sunav Choudhary. Here, we created concise user representations which can be projected onto edge server, providing faster marketing services.

Feel free to reach out for research collaboration, academic guidance (if you are an undergraduate student from BITS) or even just for a chat! For GRE and TOEFL preparation guidance, especially under time contraints do check out this repository.

Email  /  Resume  /  Google Scholar  /  Linkedin  /  Github

[1] A. A. Joshi, H. Settibhaktini and A. Chintanpalli, "Modeling Concurrent Vowel Scores Using the Time Delay Neural Network and Multitask Learning," in IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 30, pp. 2452-2459, 2022, doi: 10.1109/TASLP.2022.3192096. (Link)

[2] A. A. Joshi, P. Bhardwaj and S. M. Zafaruddin, "Terahertz Wireless Transmissions with Maximal Ratio Combining over Fluctuating Two-Ray Fading," 2022 IEEE Wireless Communications and Networking Conference (WCNC), 2022, pp. 1575-1580, doi: 10.1109/WCNC51071.2022.9771926. (Link)

S. Chakraborty, S. Choudhary, A. Sinha, S. Nair, M. Ghuhan, Y. Gagneja, A. Joshi, A. Tyagi, S. Gupta, “Generating Concise and Common User Representations for Edge Systems from Event Sequence Data Stored on Hub Systems”, US 17/849,320, Filed Jun 24, 2022. (Link)

I have been an avid practitioner and performer of Hindustani Classical Vocal Music for the past fourteen years.
Here's a short article on Raga that I had written in my sophomore year (2020). Would love to hear your thoughts!

Following is my musical journey during college and beyond:

Thanks Jon Barron for this template.