About
I've worked across pharma & clinical data, HCI research, and health insurance — and I like taking projects end-to-end, from research and design through engineering.
These days I'm focused on accessibility and vision-language models at the MadAbility Lab, shipping a medication-companion app called Pharavo, and supporting CRM data quality at TruStage.
Where I've been
Data Operations InternIncoming
Supporting Salesforce CRM data quality, reporting, and automation initiatives for Filene Research Institute.
Undergraduate Research Assistant
Evaluating vision-language models in MR/AR for an HCI accessibility project. Analyzed 500+ Meta Quest apps using a structured classification framework to surface accessibility trends; contributing to a paper in preparation.
Vice President · Product Manager
Lead operations across a 14-member exec team spanning marketing, comms, and web design. Oversee managers across 3+ local client projects each semester, supporting execution and delivery of human-centered design solutions.
AI/ML Project Intern
Engineered an end-to-end RAG pipeline using Python, NLP, and semantic search over health insurance documents. Built a Dockerized Weaviate vector DB with Hugging Face embeddings across 20+ document schemas.
Data Science Intern
Built ETL pipelines in Python and R for 10+ workflows across 3 studies, reducing audit prep time by 45%. Designed 25+ SQL-based automated validation checks; developed 6+ Tableau dashboards for cross-study monitoring.
Research Intern
Co-authored “Exploring Communal Gratitude in Online Communities” (PACMHCI / CSCW 2025). Conducted thematic analysis on user research data and synthesized findings into key behavioral patterns.
Selected Projects
Reads prescription labels with Gemini vision and offers multilingual NLP-driven translation to support adherence for ESL patients and older adults.
Pharavo
An AI medication companion built in React Native that reads prescription labels with Gemini vision and offers NLP-driven multilingual translation to support adherence.
Led a team of 8 through a Native Vote + Voting Info redesign — System Usability Score of 91, 23 points above average.
Wisconsin Conservation Voices
Led a team of 8 through a redesign of Native Vote and Voting Info pages. Achieved a System Usability Score of 91 — 23 points above average.
Production RAG pipeline + Weaviate vector DB with Hugging Face embeddings across 20+ insurance document schemas.
Spectacle Health
Vector database + semantic search over insurance documents. Delivered the production system to the client.
Built a 2,000+ record dataset across 50+ years of charts; modeled reappearance at 95% classification accuracy.
Billboard “One Hit Wonder”
Built a 2,000+ record dataset from Wikipedia + Spotify + Last.fm; modeled reappearance at 95% classification accuracy.