Scraping wastewater contaminants data from literature using large language models (LLMs) to train a machine learning model for predicting wastewater contaminant loading.

Developed a cloud-based, GenAI-powered document search engine utilizing open-source large language models (LLMs), optical character recognition (OCR), image caption generation, and vector database technologies. This system automates competitor intelligence gathering in CAR-T cell therapy. Processed study data from a Japan-based research team in accordance with a data transfer agreement (DTA).