Learning by Doing: How Professor Robert Fox Brings Real-World AI to UNC Charlotte
By Jordan Minor
Professor Robert Fox serves as a vital link between corporate AI practice and academic training at the UNC Charlotte School of Data Science. With years of experience at top companies such as IBM, Bank of America, and DataRobot, and as a current Ph.D. student here, he shares his expertise to help guide future data scientists as they begin their careers.
Bringing experience into the classroom
Before joining UNC Charlotte as a full-time professor, Fox built a long and successful career. He spent over 20 years advancing at IBM and at diverse financial institutions, always eager to keep learning.
In 2017, I decided I wanted to experience more beyond the world of large corporate financial institutions. I went to work for a startup called DataRobot, where I worked with some of the brightest data scientists on selling their AutoML platform around the world. I wanted to have the “small company” experience, but I was not one of the company’s founding members. During my time at DataRobot, I was educating others almost daily, and that’s really where my love for teaching began to shine.
Why UNC Charlotte’s School of Data Science
After earning a master’s degree from Charlotte and developing a passion for working in data science, Professor Fox felt that teaching at the University was an obvious next step.
I’ve been living in the career of a data scientist—even before it was called data science! So for me, the School of Data Science was a natural fit. Teaching has always been extremely rewarding for me…and I am continually learning along the way. Not many careers make room for more full-time learning!
The importance of model interpretability
In addition to his roles as a professor and data scientist, Fox is skilled in AI governance. In his special topics course on Model Risk Management (DSBA 6010), he highlights that model interpretability is just as important as the AI model itself.
For certain high-stakes models (like credit decisioning), model interpretability is critical, and we even need to justify individual predictions. Just because you use a model doesn’t mean you absolve yourself of accountability. A good data scientist must really know their model and help develop trust in its output. Explainability is often overlooked in basic courses on ML.
How to handle the hype
Even as a professor, Fox continues to code queries and prototype solutions to keep his technical skills sharp. With data science evolving quickly, we asked him how he filters out the hype that often comes up in everyday conversations.
That’s a great question. I’m always talking with and consulting practitioners in the real world outside academia. Practitioners build things, and hype is quickly exposed if it doesn’t work!
How storytelling shapes student perspectives
Storytelling is an important skill when sharing your work. It draws on backstory, experience, and the final product. In his Visual Analytics and Storytelling course (DSBA 5122), we asked Fox what it means for students to tell a story effectively with data visualization.
Lacking the ability to communicate effectively can absolutely hinder your ability to advance your career and get that more senior position. I realized that maybe halfway through my 20+ year career. Can you believe: 20 years ago, I was absolutely mortified at the prospect of standing in front of a classroom? And now I do it full time!
Why SDS students stand out
We enjoy asking professors what makes School of Data Science students stand out. With experience in many businesses and startups, and through teaching, Fox had high praise for the students and shared why he enjoys being a professor at UNC Charlotte.
Working hard and asking questions. That may seem simple, but it’s true. The best students show initiative and get their hands dirty, and those students always ask the best questions. Concepts and theory are great, but learning by doing is a wonderful way to really dive into anything, especially data science!