Meet new faculty member Shubhanshu Shekhar

We are delighted to welcome Shubhanshu Shekhar to the ECE community. Shekhar joined the ECE faculty this past August 2024 as an assistant professor specializing in machine learning and statistics. His research is focused on developing principled statistical and algorithmic methods for solving practical problems in machine learning and data science.

Before coming to Michigan, he was a postdoctoral researcher in the Department of Statistics and Data Science at Carnegie Mellon University (CMU). He received his PhD in Electrical Engineering from the University of California, San Diego, during which time he was awarded the Shannon Memorial Fellowship to pursue his research in information theory, and the Dr. Sassan Sheedvash Award for his research in AI and neural networks. 

He is currently teaching the course ECE 564: Estimation, Filtering, and Detection, and next term will be teaching ECE 550: Information Theory.

To help prospective students (and his colleagues) learn more about Shekhar, we asked him a few questions. 

Tell us about your research.

My research interests lie in the areas of theoretical statistics and machine learning. Broadly, I work on designing new inference and decision making methods that make few assumptions on the data source, and are valid under dynamic user behavior (such as optional stopping, adaptive sampling, etc.).  

What do you enjoy most about your field?

My research is quite interdisciplinary — it borrows tools and ideas from several fields, such as probability theory, statistics, information theory, machine learning, optimization, game theory, etc. This provides me with several opportunities to interact with and learn from experts in all these areas. 

How does your work impact the world around us?

My work contributes to the design of robust data-driven inference and decision making systems that are employed in various practical applications. As a specific example, I recently worked on designing an adaptive sampling strategy that can be used for efficient AI-assisted auditing of financial transactions. 

What’s your favorite thing about teaching? 

I like the fact that teaching gives me an opportunity to go back and relearn the basics of some topics that I had assumed I already knew. This often deepens my understanding of the subject. 

I taught a mini-course on “Information Theory and Statistics” while I was a postdoc at CMU. In the future I hope to develop and teach an expanded version of this course at the University of Michigan. 

What qualities do you look for when selecting PhD students to join your team?

I enjoy working with students who are inherently motivated about research, and are not afraid to try new ideas independently. 

What is your approach to mentoring graduate students?

While the specifics depend on the individual student, I always try to provide continuous feedback through email or slack for minor questions, and have regular one-on-one meetings to dive deeper and work through the bigger issues. 

Do you have any hobbies?

I enjoy playing all sorts of strategy games; from Football Manager to chess​​!