Co-authored a research paper studying how varying levels of self-disclosure in AI companion agents affect user perception, trust, and engagement. The study deployed a full-stack web application for structured user interactions, with a secure backend pipeline for data collection and analysis.

• Built a full-stack web application for deploying and logging structured user interactions with AI companion agents
• Designed a secure data pipeline and server compliant with university RISK protocol
• Measured user perception across multiple self-disclosure conditions using validated survey instruments
• Analyzed results using statistical methods in R to identify significant disclosure-level effects on trust and likability