Faculty Spotlight: Meet the McIntire School’s Reza Mousavi

At the University of Virginia’s McIntire School of Commerce, Associate Professor Reza Mousavi is helping shape the next generation of innovators at the intersection of technology and business. With a background spanning invention, data science, and academic research, he brings a unique perspective to his work on artificial intelligence, both in the classroom and in his research. We sat down with Professor Mousavi to learn more about his path to UVA, his teaching and research, and what excites him most about the future of his field.

Q: Can you just tell me a little bit about yourself, your role at UVA, and then the path that brought you here?

A: I’m Reza Mousavi. I’m an Associate Professor at McIntire. When I was in my undergrad, I invented a couple of things and I was really interested in finding a market for those, like commercializing and selling them, and so forth, so I decided to pursue a business degree. I did an MBA, and then I was really fascinated by how Google is transforming the world and in the intersection of technology and its business applications. That prompted me to get a PhD degree in Business Administration with a concentration in Computer Systems. During the last year of my PhD, I was headhunted to be a data scientist at State Farm, the insurance company. I was there for about a year doing research related to data science and analytics. I then decided I wanted to steer my own research direction. That desire prompted me to come back to academia, and that’s how I ended up here.

Q: What courses do you currently teach?

A: All of my courses are about AI, either their inner workings or how they’re applied to solve specific problems. I teach one undergraduate course at McIntire about the foundations of artificial intelligence. I also teach a graduate course at McIntire about using Python programming for creating AI-based models. I also teach two courses in our Master of Science in Business Analytics program (MSBA), which is a joint program between McIntire and Darden. Those two courses are about the inner workings of artificial intelligence models and cover large language models: how they’re being built, their inner workings, how you can modify them to do specific things (better), and so forth.

Q: What first inspired you to pursue your field of study? 

A: I was focused on optimization techniques and algorithms during my MBA. My MBA concentration was Operations Management, which is about optimizing business processes. Naturally, I was drawn to AI models because they’re designed to bring efficiency. We focused on how AI can automate certain things, enhance productivity, and more. That stems from my initial interest in optimizing whatever I see around me — and I’ve always been really interested in that.

Q: Are there any current projects or research that you’re most excited about right now?

A: I have many research projects at different stages, some have just come out, some are in revision for journals, and some are being written at the moment. Typically, I focus on the inner workings of generative AI models. For example, there are certain settings that you can adjust within these large language models — say, GPT or Gemini, or some of the others — and these changes affect the text they generate or the output they provide. I’m also interested in seeing how humans react to those adjusted outputs. Basically, I study the “knobs” you can turn within these generative AI models and how humans respond to them. That’s one of the streams of research I’ve pursued and plan to continue pursuing.

Another paradigm focuses on comparing how LLMs perform tasks versus how humans perform them. We have collected data by asking a group of individuals to perform certain tasks while we monitored their performance. Now we’re asking LLMs to do the same thing, and we’re comparing and contrasting. Early evidence shows that these LLMs do mimic some of the same characteristics — for example, they pay more attention to the same kinds of things humans pay attention to — so we see some convergence there.

I’m not just interested in showing that the LLMs are mimicking human “reasoning.” It’s the way they pay attention to tasks that seems to be similar, to some extent, to the way humans pay attention. This could potentially give us insights into modifying these LLMs to either outperform humans or match the performance of top-performing humans.

Q: What do you find most rewarding about teaching and being part of the UVA community?

A: I would say the students. They’re amazing; both the undergrad students we have at McIntire and the graduate students I’ve taught at McIntire and in the MSBA program. They’re really motivated to do cool, meaningful things. I think interacting with them and helping them build AI-based models is one of the most rewarding parts of my teaching. All of my students (anyone who takes my courses) is required to build AI-based applications. These are real applications that can be turned into production-ready online application others can interact with. That’s really cool, I think.

Q: What advice do you have for students or professionals considering this field?

A: First, pursue this field with the understanding that big changes are coming constantly. It’s not like you can just read a bunch of textbooks, take a few courses, and then you know what’s happening. What you know right now may not be very relevant a couple of years from now.

That said, it’s really important to understand the foundations and keep track of how changes build on those foundations. If you don’t have a solid foundation, you won’t be able to learn new things in this field. You need to have a good, solid foundation, and then you can understand how new developments essentially respond to the limitations of previous methods. It’s an ever-evolving field, and you should see that as something to pursue continuously. You should be genuinely interested in it; that’s the only way to be truly successful.

The second thing I’d say is: focus on applications and what you can do with this knowledge, rather than trying to fit into a specific mold that tech. firms expect. I encourage all my students to think like business owners. You’re equipped with these tools, you can build these applications, now go solve a problem in the world. Be open-minded about creating your own tools, your own applications, or even your own company.

Q: Have you ever visited the UVA Northern Virginia campus, and if so, what’s your favorite thing about it?

A: I really like the style and design; it’s very cool and modern. It feels like you’re in a tech firm instead of a university. The atmosphere is the most interesting thing, and I’ve met some folks there who are very friendly.