I started as a researcher. My doctoral work at the University of South Florida used reinforcement learning and game theory to model deregulated electricity markets: how generators bid, when market power shows up, and how to plan capacity under carbon rules. One of those papers, in IEEE Transactions on Power Systems, is still cited in research published today.
I then joined the University of Wisconsin-Milwaukee as a tenure-track assistant professor of industrial engineering. I taught statistics, stochastic optimization, game theory, and reinforcement learning to PhD, masters, and undergraduate students, helped found the Energy Analytics Research Group, and worked on carbon markets and the energy-water-climate nexus.
I moved to industry because I wanted to see models run. At Impetus I built distributed reinforcement learning for electricity market bidding on Spark. At IBM I led data science and engineering for IT operations analytics. At Prudential Corporation Asia I led predictive analytics and put many models into production across Prudential. At Munich Re I became Regional Head of AI and Analytics for the non-life business across Asia: I owned the AI strategy, ran the Regional Analytics Center for Southeast Asia, Japan, Korea, and India, and led the launch of an image-based motor claims product.
After leading insurance AI for Southeast Asia at Accenture, I joined Kyndryl in a global role, helping build its data and AI practice. In 2025 I moved into a regional role, leading AI and innovation for ASEAN and Korea, and opened the ASEAN AI Innovation Lab in Singapore with Google Cloud and Digital Industry Singapore.
What I believe after all of it: the model is maybe ten percent of the problem. Programs fail on data, integration, governance, and people. So I start with the outcome, design governance in early, and build for production from day one. I still teach, at NUS, because explaining something clearly is the best test of whether you understand it.
