Projects

Built and presented an autonomous underwater vehicle simulation for detecting ghost gear — lost and abandoned fishing equipment that poses serious risks to marine ecosystems. Developed in MATLAB and Simulink over 72 hours during the Courage Builder × UMass Dartmouth Blue Tech Externship, with hands-on guidance from the MathWorks team. Pitched the completed system to industry judges. The externship led directly to a summer internship offer from New Bedford Research and Robotics (NBRR) × Sedna.

Co-built an AI-powered biomedical research assistant for the Databricks Free Edition launch hackathon. Built on Databricks with Unity Catalog, the system ingests real PubMed datasets and gives researchers tools to search, summarize, cite, and analyze trends across large volumes of academic literature. Multi-agent architecture coordinates specialized tools so researchers can go from a question to citation-grounded insights in seconds.

Built an end-to-end retrieval-augmented generation pipeline for semantic search over academic PDFs. Handles the full stack: PDF ingestion with PyPDF, heuristic metadata extraction, SQLite storage, sliding-window text chunking, vector embeddings via sentence-transformers, and a FAISS index for similarity search. Natural language queries surface the most relevant chunks across a document corpus. Built as a deep dive into RAG fundamentals, SQL data storage, and data pipeline design.

Co-developed a Flutter app that gives UMass Dartmouth students personalized meal recommendations based on the live Grove Dining menu. Personally built the Python web scraper pulling structured nutritional data from the dining hall, engineered the Firebase cloud infrastructure for real-time data sync, and developed the AI recommendation model applying predictive modeling for personalized nutritional insights.