Akshay Gupte

Ella Yu and Becky Nisbet worked together to produce this article as part of our series of Academic Interviews; featuring Akshay Gupte!

I am a faculty member of the Optimization and Operational Research Group, and my research is on the mathematical and algorithmic aspects of optimization. Within that, I do problems on what’s called discrete optimization and an area called non-convex optimization.

Optimization problems are really the maths behind decision making. Most decision problems require you to choose from amongst a certain set of feasible choices governed by some constraints, and you want to either maximise profit or minimise cost or optimise the decision in some other way. These problems appear at many places in practical life, for e.g., while doing mundane things such as filling a grocery basket in a supermarket, and also in the industry across fields such as engineering, economics, finance and business. Within that, when you mathematically model these problems in real life, the some of the choices are very often discrete in nature. This means that the choices you make cannot take arbitrary values, they may have to be integers for instance. Having these choices does complicate your problems quite a bit, so there’s a lot of mathematics behind why it’s harder to solve these problems. Discrete optimization, or more generally non-convex optimization, addresses such problems.

Optimization and Operational Research is definitely not pure maths, it is in the realm of applied and computational mathematics. If I were to draw a Venn diagram to represent the field, I’d draw one circle for maths, one for computer science and one for engineering, and the field of Optimization and Operational Research would sit in the intersection of these three. In fact, my PhD is in Operational Research from Georgia Tech in the US, and was actually cross disciplinary between maths, computer science and engineering.

Can you tell me a bit about your education background or early life?

I grew up in Bombay, India and I did my undergraduate there in Industrial Engineering. It incorporated a few things from management science and applied maths.

I was always interested in, and good at, mathematics all through my primary, middle and high school years. So I knew it was something I wanted to pursue for my higher studies. At the end of high school, I successfully passed the very tough and competitive entrance exams for the prestigious Indian Institutes of Technology (IITs) which had a 3% acceptance rate from more than 100,000 examinees, and I wanted to do a degree either in Applied and Computational Mathematics, or in Aerospace Engineering because fluid dynamics in designing cars and aeroplanes was another area of interest (largely fuelled by my passion for the sport of Formula 1 racing). Unfortunately, even though I got accepted from the entrance exam, my rank was not high enough to be accepted for either of these two degrees. So, instead of compromising and pursuing another concentration area from one of the IITs, I went to the University of Bombay for Industrial Engineering.

Why did I choose Industrial Engineering? All of my friends got into software and computer engineering because that was very popular at the turn of the 21st century. But I didn’t, and Industrial Engineering was much lesser in demand. I have an uncle who did a PhD in Industrial Engineering in the US and is a full professor and department chair at a high-ranking research university. Talking to him, he said that if I do this there’s a field called Operational Research within this aspect of engineering. If Applied Maths if what you’re really interested in, then you can do Engineering for your bachelors then for postgraduate study you can specialise in Operational Research. So it was really the benefit of a close relative at a top university in the US that set my career path in motion.

So I did my Bachelors in India, then went to the US to do my MSc in Operational Research at the University of Arizona. I then did a PhD in Operational Research at Georgia Tech.

You entered this field around 20 years ago. Has there been any major changes since then?

Right around 2009 and 2010 there was this whole analytics boom where you had these big chunks of data and people wanting to use them to make decisions. Then in the last four years or so there’s been this much larger wave of big data, data science and, more generally, artificial intelligence and machine learning. This is still centrally around computer science, but statistics plays a big part in it; Operational Research has its own niche as a lot of these problems in artificial intelligence and machine learning are optimization problems.

Are you lecturing anything at the moment?

On the teaching side, at the moment I’m lecturing a course called Optimization Methods in Finance. It’s an MSc course, although some upper undergraduates take it too. This teaches how optimization can be used to solve certain problems in quantitative finance. It’s a course that’s required for our MSc degrees in Financial Maths and in Operational Research.

Do you prefer the teaching side or the research side?

I definitely like the research side of things; the majority of my time is consumed by it. Having said that I also like to teach, especially this course, and when I’m doing this I’m also learning things about quantitative finance. You get to teach and disseminate knowledge to a big population of students who are then trained, go into industry and accelerate technology even further.

Are there moments where you get inspiration from your students while you’re teaching?

I get lots of good questions from them when I’m teaching and sometimes they stimulate my mind and inspire me to think about other research questions.

Can you give some suggestions to maths students about how to find their area in maths?

Pursue your passions, but take time to gauge your interests. It doesn’t necessarily have to be when you enter University; it can be at the end of year one or two, but definitely try to do well in your courses. Try to pick something that you’re passionate, and interested, about and that you’ll be able to contribute and make an impression in. As I can tell from my own life trajectory, you do not always have to follow what is in vogue or jump on the bandwagon to be successful in your profession.