91´«Ã½

Can you tell us about your educational and work experience background before you enrolled in this program?Ìý

I have a BA in English with a math minor, a BS in electrical engineering, and an MFA in creative writing.Ìý

What initially drew you to this program?Ìý

I've had a long career in the tech industry, including 22 years at Microsoft. Over that time, I've come to appreciate the importance of first principles in data science: understanding its statistical foundations, the intuition behind algorithms, and the motivations that guide their design and use. I began looking for an academic program that would deepen my theoretical knowledge while broadening my technical skills. UC Boulder’s fully online, accredited program proved to be the perfect fit.Ìý

Can you tell us how the MS-DS program fits into your life?Ìý

My personal and professional commitments would make it difficult for me to attend classes in a traditional academic setting. Because CU's MS-DS classes are all online--and self-paced--I can fit them into the odd times of day when I'm free: I can listen to lectures on long runs, complete the coursework in the early mornings, and catch up on discussion forums in the evenings.Ìý

What are your favorite parts of the program?Ìý

My favorite part of the program has been the labs, because they're when I get to apply the concepts that were introduced in the lectures. They let me put theory into practice, deepen my understanding, and experiment with ideas in a hands-on way. And by the time that I finish a lab, I often have a much better grasp of the material and a much greater appreciation of its value, which is especially gratifying.Ìý

What do you hope to do with your MS-DS degree?Ìý

Data science is a rapidly evolving field—new models, tools, and technologies are constantly emerging—so in many ways we're never finished learning. Fortunately, the MS-DS program has given me the skills and confidence to continue growing with the field. I now feel much more comfortable digging into the technical details of algorithms and working through the formulas in academic papers. In truth, the program has helped me push beyond limits I once placed on myself, and I plan to use my degree as a foundation for continued learning.Ìý

Would you recommend this program to others? Why or why not?Ìý

CU's MS-DS is a great fit for anyone who's looking for a rigorous program but needs the flexibility of an online learning environment. They'll need to take ownership of their own learning, but if they're self-motivated and disciplined, they'll find the program both challenging and rewarding.Ìý

What do you wish you’d known before starting the MS-DS?Ìý

If anything, I wish that I had discovered the program sooner.Ìý

What’s one tip you have for students who are starting this program?Ìý

Don't presume that the courses will be taught the same way. Each of the instructors brings their own personality and their own pedagogical style to the courses they teach. Embrace those differences—they're part of what makes the program so engaging.Ìý

Is there a specific project you have worked on that stands out to you?Ìý

I always appreciate when labs give us the freedom to define our own projects. I always try to come up with something I'm truly curious about—a question I genuinely want to answer. For me, that reinforces one of the most practical aspects of the program: the ability to apply what we’ve learned in real-world scenarios. In my case, the questions that I'm curious about often come from my work at IQRush, where we focus on AI visibility. AI--particularly LLMs--represents a huge inflection point for many industries, and one of the exciting things about this space is how many unanswered questions there are. Increasingly, people are using generative search to answer questions and discover information online. Naturally, brands want to know how often their products or websites appear in those answers. Several companies now offer services that attempt to measure AI visibility. Typically, they ask the generative search engines a set of queries and count how often a brand's products or links appear in responses. The results are often presented as if they were deterministic measures of a brand's visibility. In reality, they're only estimates. The whole process contains a substantial amount of variability—from the specific questions being asked to the stochastic nature of LLMs themselves. Ask the same question twice, and you'll get different answers, so if you repeat the same measurement process, you will get different results. But how different? That simple question led to my paper, "Quantifying Uncertainty in AI Visibility," which explores how statistical methods can be used to measure and interpret that variability. Fortunately, we learned how to address problems like these in our coursework. In this case, the Statistical Inference specialization was especially helpful, but I also applied skills from several other specializations: Vital Skills for Data Scientists, Data Mining Foundations and Practice, Data Science Methods for Quality Improvement, and more. I've already posted a preprint of the paper on arXiv, but when I submitted it for publication, I was especially proud to affiliate myself with the University of Colorado Boulder.Ìý