Samarth Gupta

 

Samarth Gupta
Applied Scientist 2 at Amazon
PhD, Department of Electrical and Computer Engineering
Carnegie Mellon University

Email: sidsamarth@gmail.com
LinkedIn
Google Scholar

About Me

I currently work as an Applied Scientist at Amazon where I am working on recommendation systems in the personalization org. I received a PhD from Carnegie Mellon University, where I was advised by Prof. Gauri Joshi and Prof. Osman Yağan. My Thesis "Structured and Correlated Multi-Armed Bandits: Algorithms, Theory and Applications", studied sequential learning algorithms in presence of noisy and correlated/structured observations. I like to work on problems involving Statistics, Machine Learning, Optimization and Performance modeling. I am interested about the applications of these problems in A/B testing, experiment design, recommendation systems, model selection in ML, robust machine learning and metric design. At Microsoft, I worked on productization of LLMs for Windows co-pilot, new features in Bing Chat, text prediction in Edge browser (EN and non-EN languages). I have also been involved in inference latency optimization of LLMs at Microsoft.

At CMU, I explored problems in online learning and statistical inference. I designed and analysed online learning algorithms to solve sequential decision making problems. In my PhD, I did some work on Correlated Multi-Armed Bandits (1, 2, 3), Stuctured Multi-Armed Bandits (4) and Active Distribution Learning (5). In these works, I have shown the use of our proposed algorithms for the application of recommendation systems and combinatorial online resource allocation problems (6).

During Summer 2019, I worked as an intern with the prediction team at Uber ATG to develop uncertainty aware metrics and safety-oriented predictions for autonomous vehicle. In summer 2021, I worked as an intern with the search team at Amazon, where I studied the problem of selecting best among the available search engines for the query typed in by the customer using session-aware MDP based models (7).

During my undergrad at IIT Bombay, I had a chance to work on resource allocation problems for large scale content delivery networks (8,9) with Prof. Sharayu Moharir .

Here is a link to my CV.

Aside from research, I love to run, bike and play tennis, squash, table tennis and soccer. I try to follow tennis and soccer quite regularly.