Data Scientist
Current- Developed a Repository-Relevant-Information-Collector (RRIC-AI) tool, significantly reducing manual search time by over 500 hours for Nuclear Engineers and other Subject Matter Experts- Automated the migration of 50+ years of unstructured Nuclear Decommissioning data stored in PDFs and PowerPoints from the Waste Management Symposia website using a Python Selenium script, saving 200 hours of manual work- Applied advanced NLP techniques to OCR, clean, assure quality, and semantically chunk a diverse corpus of over 13,000 files, optimizing chunking strategies using LangChain and LlamaIndex- Employed BERT embeddings model to encode chunked data into vectors, creating a robust vector database for enhanced semantic information retrieval