Experience

Developed AI-powered Q&A chatbot using open-source large language models (Phi-4, Llama-3, Gemma-2B) to provide farmers streamlined access to agricultural seed laws across 78 countries Processed and analyzed 183 legal PDF documents across 9 languages and covering legislation from 1981-2023 Built end-to-end Retrieval-Augmented Generation (RAG) pipeline with ChromaDB vector database, implementing semantic search and document chunking strategies for legal documents ranging from 1-150 pages Collaborated with A Growing Culture nonprofit organization to advance global food sovereignty through multilingualdocument processing system development Optimized chatbot response time to 15-20 seconds and evaluated model performance using ROUGE metrics across multiple LLM architectures Identified critical gaps in multilingual NLP preprocessing pipelines and developed custom text normalization strategies for non-English legal terminology across 9 languages

Developed comprehensive Census data dashboard to replace complex Census Bureau website navigation, improving researcher access to demographic insights Automated processing and analysis of over 100,000 U.S. Census records spanning 2009-2023 with real-time filtering and visualization capabilities Created full-stack interactive dashboard using Python, Shiny, and Pandas, resulting in 40% increase in stakeholder engagement and 35% improvement in user experience metrics Presented research findings and demographic insights to audiences of 200+ attendees at research symposiums Translated complex census data into accessible findings for community organizations, policymakers, and academic researchers