At ADP, I built and shipped LLM evaluation frameworks for GPT-4 systems and RAG-based Q&A pipelines over enterprise transcripts. My evaluation work improved grounding precision by 30% and helped bring production insights to users.
At PlainText Lab, Penn State University, I fine-tuned transformer models on privacy policies and built explainability pipelines that show how models infer meaning, risk, and consent. I also designed a user study connecting model reasoning with how people read and understand digital policies.
My research has resulted in 10 peer-reviewed publications on explainable and responsible AI, privacy policy intelligence, and human-AI interaction. Earlier at ADP, I delivered the One Login Page authentication platform using ReactJS, Java, and Spring Boot.
