Calculated Conversations: Data Science as a Team Sport: with Prof. Kanshukan Rajaratnam (Stellenbosch University)

Professor Rajaratnam discusses how his industry experience shapes data science education, the unique interdisciplinary approach of Stellenbosch University's BDatSc degree, and why aspiring data scientists must focus on collaboration and soft skills.

South Africa is quickly becoming a hub for data science talent, with the University of Stellenbosch leading the way. Data science is not just about algorithms and models. It is about solving real-world problems. Few understand this better than Professor Kanshukan Rajaratnam, whose career spans both academia and industry. As the Director of the School for Data Science and Computational Thinking at Stellenbosch University, he has seen firsthand how business challenges shape research and how education prepares future data scientists. In this conversation, he shares his insights on leadership, learning, and South Africa’s unique approach to data science.

Your career spans both academia and industry. How has working in business influenced the way you approach education and research?

My doctoral studies and my subsequent research topics in the immediate subsequent years were directly answering questions that I faced while working in industry. My research focus was in decision making in the context of loan portfolio and many of the research questions were from the perspective of a loan portfolio manager trying to maximize profits and other metrics (such as market share). My experience in industry was also instrumental in generating research questions from problems faced by practitioners. Perhaps most importantly, working in industry gave me leadership skills that helps me in my work up to today. In industry, one is faced with diverse stakeholders and one must learn to…

Beyond technical expertise, what is one way you help students think differently about data science?

It is important to get one’s hands dirty in data science. There are many opportunities to solve problems such as through hackathons and through platforms like Zindi. Potential data scientists must get experience in building models and solving problems through these means. Additionally, one should not be a data scientist in isolation but be able to work in diverse teams, particularly with domain experts. Each domain area has nuances that one does not readily know of as a data scientist, but experts in the domain area have experience in identifying and dealing with such nuances.