About me: My work combines applied modeling with production software: designing algorithms, building experimentation frameworks, and deploying systems that translate data into decisions. At Amazon Alexa, I developed reinforcement learning models that dynamically tuned search parameters in the speech-to-text pipeline, reducing latency by 30%.

I’ve also built demand estimation models used for transportation policy analysis, production data validation systems supporting large-scale software launches at Amazon, and experimental platforms for sequential decision-making problems involving multiple interacting interventions.

After completing doctoral coursework and earning a master’s in Operations Research at Columbia, I’m interested in building systems that turn complex models and data into reliable tools for understanding patterns and guiding decisions.