I am an undergraduate researcher specializing in behavioral economics with quantitative methods, combining data-driven insights with human-centered perspectives to solve complex strategic problems. My work integrates experimental design, statistical analysis, and cross-cultural insights to understand decision-making processes in organizational and policy contexts.
This research proposal explores how sequential evaluation environments distort not just subjective impressions but also categorical and numerical candidate metrics. It introduces a behavioral experiment and formal modeling framework to test whether anchoring and trait salience systematically bias structured resume assessments. By showing how even seemingly objective traits are subject to contextual distortion, the project aims to challenge conventional assumptions about fairness and accuracy in high-stakes decision-making.
Designed and programmed three oTree experiments investigating market unraveling dynamics in climate insurance markets. Built comprehensive front-end and back-end experimental infrastructure with randomized role assignment, real-time data capture, and adaptive participant grouping. Research contributes to PhD dissertation integrating laboratory and field data from Colombian farmers, examining behavioral responses to climate risk.
I designed and built UniSearch to help students discover and compare over 1,500 universities worldwide, especially those applying to study abroad programs like Erasmus. The platform integrates a dynamic search engine, personalized recommendation system, and real-time API health monitoring—all wrapped in a responsive, clean UI. This project showcases my ability to connect frontend and backend systems using React/Vite and FastAPI, implement user-centric features, and deploy full-stack apps across Vercel and Render with proper CORS configuration.
I built the Behavioral Finance GPT Assistant to classify investors not only by how they see themselves, but by how their responses reveal hidden cognitive biases. The assistant infers likely behavioral pitfalls—such as overconfidence, loss aversion, or inertia—then generates mapped investment plans tailored to each user’s profile. It pairs language-based archetyping with structured portfolio recommendations, integrating live financial data to ground advice in current market conditions. This project demonstrates how behavioral theory and real-world analytics can work together to deliver adaptive, bias-aware investment strategies through a conversational interface.