Luke Butler

- Architected a full-stack RAG platform (Python, Flask, Gemini API) that semantically searches 500+ student essays and generates analytical responses, the first LLM of its kind at UNCW. - Designed and implemented a complete ETL pipeline to ingest, parse, and embed unstructured documents using Gemini's API's embedding model, with a JSON caching layer enabling 14ms retrieval. - Engineered a semantic retrieval system using 768 dimensional dot-product similarity via Numpy, to match user queries to student papers by meaning and emotional content rather than keyword overlap alone. - Drove system architecture from concept to deployment across 20+ prototype iterations, translating abstract research goals from stakeholders into a concrete roadmap & final product.

- Co-authored a peer-reviewed research article published in Physical Activity and Health (2025), investigating culturally tailored health interventions for high-risk populations. - Managed the primary recruitment phase of the 'Shero' study, utilizing convenience and snowball sampling to successfully onboard 50+ participants. - Conducted eligibility screening and demographic data collection, ensuring all participants met strict clinical criteria regarding family history of breast cancer. - Executed research protocols with 100% data integrity, including survey administration, interview logistics, and contributing as a trained notetaker during qualitative focus groups.