Welcome to my website! I am an Assistant Professor in the Engineering Systems & Design pillar of SUTD. 

My research interests lie in analytics & operations research, with a focus on how to transform data into large-scale planning and operational decisions which are efficient, robust, and reliable in the face of uncertainty. Our research has been contributing to building data-driven optimization models in which applied probability tools such as optimal transport, large deviations, and limit theorems are used in a novel manner to tackle fundamental challenges relating to robustness and risk in decision-making under uncertainty. These contributions have been recognized with the biennial INFORMS Applied Probability Society best publication award (2023), the INFORMS Junior Faculty JFIG paper competition Third Prize (2021), and Winter Simulation Conference best paper award (2019). My CV can be viewed here: link to CV

I currently serve as an Associate Editor for the journals Operations Research, Stochastic Systems, and Operations Research Letters.  

Methodological research areas: Applied probability, data-driven optimization, Monte Carlo methods for optimization under uncertainty, machine learning

Specific topics: Distributionally robust optimization, rare events analysis, and large deviations

A specific focus of my recent research has been on learning to optimize under data imbalance and rarity, a term referring to situations where only a small portion of the dataset has an outsized impact on estimating  quantities consequential for decision-making. Such an imbalance due to rarity is witnessed all too often, with some relatable examples being 

1. quantitative risk management: eg, loans with defaults comprise only a small fraction of a loan dataset;

2. online retail: eg., a large fraction of  products have zero or low sales;

3. machine learning: eg: few object categories have numerous image examples & numerous object categories have very few image examples;

4. managing power grids: data relating to black-outs is negligible in power transmission datasets. 

Failing to tackle this imbalance and scarcity in relevant data can lead to poor operational outcomes in terms of  risk, safety, or fairness.  Our goal has been to understand and bring out how one can tackle this challenge by allowing a transfer of knowledge from "relevant, relatively data-rich" portions of the dataset to the data-scarce tail portions in a systematic fashion and how it can benefit decision-making under uncertainty. 

I am also enthusiastic about working with the industry on tackling challenges which require novel analytics and risk management solutions.