Introduction:
Professor Seungki Min graduated from the Department of Electrical and Computer Engineering at Seoul National University and earned his Ph.D. in Business from Columbia University. He taught at the Department of Industrial and Systems Engineering at KAIST for three years and joined the SNU Business School in the spring of 2025. He researches automated decision-making algorithms and teaches the Management Science course.


ACADEMIC JOURNEY AND RESEARCH PHILOSOPHY

Q. What inspired you to switch your major from electrical engineering to business?
After working for three years as a skilled industrial worker at a finance-related IT company, I spent about three years as a quant developer developing algorithmic trading strategies and programs for real-time financial data processing and automated decision-making. While I enjoyed my job, I decided to study abroad to deepen my understanding of the decision-making process. I initially considered studying control theory in electrical engineering, but as I prepared to apply to universities, I discovered active research on algorithmic trading was being conducted in a branch of business administration called operations research. I became fascinated and decided to pursue a Ph.D. in business. Though it may seem like a drastic career change, it was a natural decision for me. My Ph.D. advisors also majored in electrical engineering, and I incorporate many methods from electrical engineering, such as information theory and control theory, into my research.

Q. As a researcher, what do you think is the most interesting or intriguing question in operations management?
The question I am most interested in is, "How do data-driven decision-making models learn?" Machine learning and AI techniques have advanced significantly in recent years, and their ability to handle highly complex and unstructured prediction/generation tasks is unparalleled. However, I believe the ultimate goal in operations management is decision-making rather than predicting or generating results. Although it is important to employ advanced prediction/generation techniques in decision-making to achieve better outcomes than traditional models, my interest lies in exploring the fundamental gap between learning prediction/generation models and decision-making models. For instance, this gap can occur when the data collection process itself is part of the decision-making or when the evaluation metrics of the prediction/generation model and the decision-making model do not match. I am interested in confirming the existence of this gap and developing methods to address it.

Q. How are your experiences as a quant developer and academic researcher connected, and what do you think are the differences and synergies between these two areas?
In the field, you aim for profit and must create something that works. In academia, you pursue the truth and need to refine your understanding and insight. The industry and academia pursue different values and operate differently. However, the way they advance can be complementary. Research far removed from reality can be far-fetched, and practice without principles has the danger of becoming haphazard. Each area can learn from one another.
When I conduct research, I become curious about what the practitioners are thinking. Is my research conveying a realistic message? What are the current concerns in the field? Conversely, when you work in the industry, you find yourself wondering what researchers are thinking. Is there a correct answer to the question I'm considering? Is there a more up-to-date method? I believe theory and practice are symbiotic.
In terms of synergy, I think the Business School offers a very special system that enables practitioners and scholars to engage with each other. I also look forward to hearing opinions and stories from many people in the industry.



MAIN RESEARCH QUESTIONS AND INDUSTRIAL APPLICATION

Q. How can your research topics― "sequential decision-making" and "bandit algorithm"―be applied to solving business problems?
My research focuses on algorithms that make decisions actively while processing data in real-time and adapting to changing environments. I am interested in applying them to algorithmic trading, online advertising, product recommendations, and active learning of customer preferences. I develop and analyze algorithms that determine how to handle large orders based on market conditions, which products to recommend using customers' personal information and past activities, and which questions to ask to understand customer preferences.
Since many service industries have shifted to online automation, algorithm-based sequential decision-making will become even more important. "Bandit" is a branch of decision-making that concentrates on gathering data and developing models, which can be used to learn more about products or customers through experimental decision-making in a situation where new products or customers keep emerging. If you are recommended a video on YouTube with only a few dozen views, the bandit algorithm may have been in play.

Q. What are some of the limitations or precautions of data-driven decision-making, such as customer preference learning or algorithmic trading?
I am sorry to disappoint you, but data-driven decision-making does not work miracles. For example, it will not produce brilliant decisions humans cannot make. I believe its significance is that automating the decision-making process can reduce overall costs and systematize the process. While it can help make many decisions quickly with the same quality, I do not think the quality of individual decisions will be better than the careful judgment of a human expert. The greatest strength of algorithmic trading strategies is their ability to respond in microseconds to market changes, but they will not be able to read market trends better than human traders. Perhaps this will change in decades as AI technology continues to advance, but for now, the limitations are clear.

Q. Could you share any recent research achievements that made a strong impression on you or topics you would like to explore more in the future?
Since I have been focusing on theoretical research, I am afraid I do not have many outcomes that might be relatable to everyone. For example, the recent work I am most proud of is proving that the "difficulty of nonstationary bandit learning problems is determined by the entropy rate of the environment," which I cannot explain in simpler terms.
On a lighter note, I have been working with undergraduate students to develop and analyze optimal strategies for "Yutnori" through reinforcement learning. We found that casting first has a slight advantage of 0.1%, and my current chance of winning can be roughly estimated as: 50% + (number of marks I brought home - number of opponent's marks they brought home) x 10%.
As I mentioned above, I plan to continue researching methods to reduce the gap between advanced AI techniques and decision-making.


EDUCATIONAL PHILOSOPHY AND TEACHING EXPERIENCES
Q. What are your most important goals when teaching courses on operations management, artificial intelligence, and data-driven decision-making?
I want my lectures to make students think. In an age with many excellent online teaching resources and convenient learning tools like ChatGPT, I often wonder what the desirable role of university education could be. Instead of focusing on transferring advanced knowledge, I try to engage with students face-to-face and encourage them to take some time to think on their own during class. It's not an easy task, but I want them to go through the process of finding an answer rather than just telling them the correct answer. I want to enhance their ability to avoid panicking when faced with unfamiliar problems and to work carefully through them.


CAREER ADVICE AND REFLECTIONS

Q. You have experienced the different worlds of engineering, finance, and academia. Why do you believe it's crucial for students to have such diverse experiences, and how can it broaden their perspectives?
I am a bit cautious because I think the situation can be different for everyone. In my case, I had many different experiences related to algorithmic and data-driven decision-making, which gave me a somewhat unique perspective that combines computer science, electrical engineering, and financial engineering. I do not believe the number of experiences alone can be helpful; rather, it is important to consider the same question in different settings and viewpoints.

Q. Is there any experience from your undergraduate or PhD program where you thought, "I am glad I did this?"
I am glad I had the work experience because it not only helped my research but also cleared my illusions and regrets about not getting a job or starting a business. Without that experience, I might not have been able to focus on my research as fully during my Ph.D. program because of the distraction of the prospect of finding a job or starting a business.

Q. What career advice would you give to students in a time when the lines between data science, AI, finance, and operations management are blurring?
We live in a time without a clear answer. This might sound cliche, but try many things, fail often, and grow a lot. I like the quote by movie critic Dong-jin Lee, "Live every day to the fullest, and let your life unfold as it may." While I am a researcher of rational decision-making, I believe that forcing rationality on the big decisions in life, such as your career or marriage, is a recipe for disaster. Don't be overly calculating and trust your heart. Make every day count. You will discover your true self along your life's journey, and this will become the fuel for your life.

Q. Do you have any advice for students who want to gain experience in both industry and academia, like you did?
While I recommend exploring both areas these days, I would advise against using them as an escape from reality or to justify yourself. When you work, you miss school; when you study, you miss working. Although there is no need to be overly calculating, there is no point in chasing fantasies.


PERSONAL VALUES AND VISION

Q. What are the philosophies and principles you follow in your research and teaching?
I do not have a grand philosophy or principle yet, but if I had one, it would be to research and teach about things that interest me. I research and teach, believing that others will enjoy it only when I enjoy it.

Q. What are your tips for managing stress and recharging yourself amid your busy schedules?
I am also inexperienced with stress and time management. The things that have worked the best these days were maintaining a healthy diet and spending quality time with my family. When I feel stuck in my research, going for a walk always helps.

Q. Do you have any academic visions or social contributions you would like to make in the future?
In the age of AI, I hope to establish the identity of my academic field, Operations Research. I want to become a producer of knowledge rather than just a consumer of AI. Furthermore, as a member of the Business School, I hope to reduce the barriers to mathematical and technical methods for undergraduate and MBA students who lack a background in science. Although I do not have a specific plan yet, I would like to contribute to the change that the Business School will lead through my courses, lectures, publications, and podcasts.