Democratizing the deployment and use of artificial intelligence: Meet new professor Rafid Mahmood

Professor Rafid Mahmood joined the Telfer School of Management in summer 2022. He’s an assistant professor in business analytics and information systems whose research focuses on applications in health-care operations and large-scale artificial intelligence (AI) systems. Before joining Telfer, he worked as a research scientist at the NVIDIA Toronto AI Lab. We interviewed him to learn more about his research interests.

Why did you choose to study business analytics and information systems?

During my studies, I moved around between engineering, business analytics and computer science, with the goal of finding connections between different fields. 
For example, I worked with medical physicists on using AI to create personalized cancer treatments where I’d say, “Hey, this is actually a generative imaging problem that is well-studied in computer vision.” I’ve also worked with technology companies where I realized, “Hey, our decisions around data collection can be optimized using techniques from inventory management.”

How does your PhD training inform your current research program?

Early in my PhD, I participated in two analytics hackathons, one hosted by a quantitative finance firm and another by the NBA. I saw people build creative AI solutions, such as estimating a company’s stock price from big data or quantifying an athlete’s skillset from video data. This got me thinking about how we can use AI to improve operational decision-making, which eventually got me interested in finding out how operations research might improve the practice of AI development.

Do you have any new research highlights to share?

I am excited to share the news about two recent studies.  In the first one, my research team and I explored a problem faced by anyone who has tried to implement an AI model, that is, how much data does their model need to perform at the level that they want? Our research allowed us to develop a formal approach for making accurate decisions and avoiding over/under collecting data. In the second paper, we worked with a milk bank in Toronto to build a prescriptive model that automated their donation system. This was exciting because when we implemented it at the clinic, we found it significantly improved their operational practice.

What kind of impact can your research have on business communities and beyond?

I see my research affecting anyone looking to develop AI systems, from technology companies building large-scale AI systems to businesses interested in using data to improve their operations. Over 95% of companies developing AI projects are challenged by data quality and quantity. This motivates my research into better managing the data collection process in AI development. At the same time, I’m always looking to collaborate with stakeholders to see how we can build impactful AI solutions to their operational problems.