When I started the MSQM: Business Analytics program, I was living abroad and looking to pivot into a new, data-driven career. I was looking to gain a solid technical foundation, but ultimately found much more than that. MSQM introduced me to a new way of thinking about open-ended problems and navigating ambiguity with confidence. Below, I share a few reflections on how the program has shaped my approach to problem-solving and my transition into AI.

1. What motivated you to pursue the MSQM program at Fuqua, and what skills were you hoping to build when you started?

I began the program when I was living in Beijing, China. I was working as an elementary school teacher at the time, and I wanted to transition to a more technical role in conjunction with a planned move back to America. Having previously worked in investment banking and film financing, I had some software skills, but they didn’t go beyond Excel spreadsheets and Bloomberg terminals.

As part of going from wrangling kids to wrangling data, I wanted to gain skills like programming, machine learning, and large language models (LLM). I chose the program because I wasn’t looking to just read a bunch of proofs and memorize some formulas, but rather learn how these tools can be used in the realm of business and beyond…  don’t get me wrong, though, I love memorizing formulas too!

2. Can you share an example of a hands-on project you’ve worked on in the program and what you learned from that experience?

I found the team projects in our Programming for Data Analytics course to be particularly valuable. We had a large set of medical data and needed to answer some commercial questions. There are tons of “teach yourself to code” courses out there, but programming for analytics is a distinct skill.

Instead of having a clear problem for which you are trying to find a clever solution, you have an open-ended goal and many plausible tools. Figuring out what techniques to use is the tough part; successfully implementing them is almost secondary.

On top of a new approach to problem-solving, working in a team was a big part of the experience for me. It made me feel connected with my teammates even though we were scattered all over the world, and it encouraged me to approach problems with flexibility, open-mindedness, and humility.

3. Your Quantitative Management Data Competition project focused on using an LLM to analyze Chinese names. How did you approach that problem? What made the project meaningful or challenging?

My project was inspired by the richness of the 18^th-century Chinese novel, The Dream of the Red Chamber. It has dense references to everything from Buddhism to imperial politics to tea, and the cast of characters is so huge that it is a challenge for even the motivated reader to keep track of them. I thought it would be interesting to see how an LLM could help with that task.

First, I needed to quantify the relationships between characters. I did this by measuring how often they are mentioned together in the text. The obvious problem is that characters in novels are often referred to by pronouns. I used BERT, a now old-fashioned model that was partially trained to guess hidden words in sentences. It pored through the entire novel, guessing which character corresponded to each pronoun. I then ranked the characters based on their textual closeness.

The most challenging part was figuring out how to turn the results into a meaningful graph. Network theory turned out to be a lot more complicated than I expected. But that’s how these things go: the part you thought would be the simplest turns out to require the most time.

4. How are you already applying what you’ve learned in MSQM — either in your current role as an AI engineer or in how you think about solving problems?

I recently started a new role as an artificial intelligence engineer at a construction conglomerate in Chicago. It’s a small team that oversees data science and automation work for a large group of companies. We do things like develop tools so that field workers can query project data using natural language instead of clunky Excel files.

Some skills I use at work are those I learned in the program, while some are not. For most of the big problems we come across, it’s not clear what tool to even use. That is where I find the most value in what the program has taught me: the ability to analyze problems, choose the best option among a range of imperfect tools, and figure out how to put it into practice, even if it’s outside your comfort zone.

More than anything, I hope this spirit of problem-solving that the MSQM program has taught me will be something I carry with me for the rest of my career.