Marquette Business AI Innovator: How a Marquette Business Graduate is Shaping the Future of Investment
When Andrew Mizwicki was building his first chatbots, he tested them with a simple question: “How many kangaroos are in Australia?” If the model gave any answer, Mizwicki knew it was off track. “The models I was building were meant to be subject matter experts,” he explains. “If a model analyzing Home Depot’s financial data answers a kangaroo question, it’s including irrelevant information.”
Now a junior quantitative associate at multinational investment firm William Blair, Mizwicki applies the AI skills he gained from Marquette’s AIM Program to help expand the firm’s AI capabilities.
How did you first get interested in building AI models?
My introduction came early in the AIM program when Dr. Joe Wall showed me how AI works. He explained how AI can boost coding skills and deepen research insights. At first, I was skeptical about the accuracy of these models, unsure if they might just fabricate information. But once I started using them, their practical applications became clear.
Can you walk us through how you built specialized chatbots in the AIM program?
AI models are trained on data. ChatGPT, for example, is trained on vast internet data. In AIM, we built smaller-scale versions focused on specific subjects. For a Home Depot expert bot, I gathered detailed financial data—revenue, costs, operating expenses. The process involves collecting comprehensive data, choosing the right libraries and coding languages, then developing and testing the model with various prompts to ensure accuracy.
How does William Blair approach AI differently from other firms?
At William Blair, we focus on practical value. We ask:
- Can AI save us time?
- Are there unique research perspectives or real-time data we can access?
- Can AI help us detect risks or exposures earlier?
It’s not about using AI just because it’s trendy. The goal is to add value across research, risk management, and performance.
What’s the most important factor in building a good AI model or tool?
Data quality is everything. You can have complex models or detailed prompts, but without enough clean, well-managed data, accuracy suffers. That emphasis on clean, understandable data was a key focus in the AIM program.
How did Marquette prepare you for your current role?
Marquette offered strong support and easy access to professors. I spent a lot of time in Dr. Wall’s and Hunter Sandidge’s office hours, and I reached out to other professors for guidance. They helped me apply classroom principles to real-world challenges.
For those interested in AI in finance, exploring targeted training can be a great next step. Check out AI tools tailored for finance professionals or browse courses designed by job role to sharpen your skills.
Your membership also unlocks:
