CSE student finds niche in research and podcasting

Tanmay Agarwal talks about his interests in computer science and AI

As a computer science major, Tanmay Agarwal believes that pursuing a career in research should be just as lauded as getting a job at big tech companies like Google or Microsoft. So, the College of Science and Engineering student started his own podcast—“Undergraduate AI,” where he interviews professors and industry researchers.

The podcast gives a “behind-the-curtain” look at how the researchers entered their careers, how their failures have guided their successes, and why they love their jobs.

“I wanted to create a resource that could help someone in their early years at college look into this world of research and be able to consider it as a highly rewarding career opportunity,” explained Agarwal, a recipient of the CSE Hopper-Dean Scholarship Honoring Dr. Vipin Kumar and U of M Scholarly Excellence in Equity and Diversity (SEED) Award.

While his initial goal was to provide advice to other undergrads, Agarwal found that interviewing professors helped him grow as a student and researcher as well. He started the podcast in summer 2020. However, Agarwal had to put it on hold once school began in the fall. He’s graduating in Spring 2021 and plans to revive “Undergraduate AI” while working as a computer vision engineer at Sentera, an agricultural tech startup in Minneapolis—a full-time position he landed recently.

In this Q&A, Agarwal discusses his passion for artificial intelligence and machine learning and his experiences with internships, student groups, and undergraduate research in CSE.  

Why did you choose to study computer science at the University of Minnesota?

I have always had an ambition to build products or services that can impact a person's daily life. I was originally studying physics and quickly realized that the major lacked the consumer-centric aspect that I'm so passionate about. I then realized the multifaceted impact AI technologies could have, and how computer vision has applications in industries ranging from agriculture and manufacturing to healthcare and surgical robotics.

This was the first “trigger” for me towards computer science. I found something that was intellectually challenging with a huge scope of research to explore that could also be leverage to help people and improve our society.

What specifically drew you to artificial intelligence and machine learning?

I am passionate about developing technologies that mimic human cognition. For example, one could use mathematical and statistical models to model and mimic how the human brain would interpret videos or signals received by the eyes, and then try to model it artificially for robots. I think this work will be very fulfilling and have positive results if done right. There are also lots of potential applications in fields like manufacturing, life sciences, and consumer robotics.

What inspired you to start your own podcast “Undergraduate AI?”

I think a lot of times—especially in the tech industry—we focus on jobs in the industry so much that we don't talk enough about how research and careers in research are meaningful. As a senior in computer science, I personally had so much resources and support around what to expect from the industry that prepared me really well for it, but I rarely found anything that taught me how to excel at research.

It can be very easy to think about pathways to "landing a job in big tech,” but I felt that the stories of the people that truly make the incremental innovations that move the world ahead ought to be shared with the same vigor.

I was a part of this lab where I was surrounded by brilliant, kind, and creative people. I felt that a lot of other students in my major might not have even considered research as a viable option. I began interviewing professors at the University of Minnesota, and my first episode was actually on my research professor Junaed Sattar! This experience has made me realize how innovative research can truly be. Academia and research allow you to explore new fields and be continuously innovative in your work, and that's something that I personally find really motivating.

Tell us more about your research with the Interactive Robots and Vision Lab.

The research team, led by [assistant professor of computer science] Junaed Sattar, works almost exclusively on underwater robots.

I'm working to develop perception-based deep learning algorithms for underwater human-robot collaboration. My specific research involves predicting diver motion to improve diver following by robots.

These robots are meant to help a diver who might be swimming for some conservation tasks or exploring the terrains of lakes and oceans. In order for robots to assist humans with these tasks, they need to follow these scuba divers in a particular direction. However, the robots are currently unable to detect abrupt directional changes, such as turning left or right. I'm also tackling how to program the robot to lead the divers instead of just following them—this is something that is completely unexplored. The hope is to determine an algorithm where the robot is able to predict where the diver will swim to next, for example, 5 seconds ahead.

You've participated in quite a few internships. What has been your favorite experience?

Each internship has served a different purpose for me because my career path has been very wavy so far. I think the most important internship for me would be at Starkey Hearing Technologies. The project I worked on was to develop a robotic arm that moves a hearing aid so that it produces the same motion signature as a human walking. This robotic arm would then help engineers to develop algorithms for the earpiece. I went into this position as a physics major with zero computer science experience (although I was taking online courses in robotics and AI at the time).

What advice do you have for incoming computer science students?

I think computer science is currently a field that tends to get a bit crowded, especially in AI and robotics. Don't let this intimidate will be all right! There will always be people who end up doing really well really fast. However, it is important to remember that everyone learns at a different pace and grows at their own speed. Don’t judge yourself based on others and don’t be scared of others’ accomplishments, especially when you’re just getting started with the program.

Listen to the first two episodes of Undergraduate AI.

Edited with permission from the Department of Computer Science website

If you’d like to support students at the University of Minnesota College of Science and Engineering, visit our CSE giving page.