Oversaw the development and implementation of AI services in the company. Responsible for familiarizing employees with LLM engineering and developing secure, reliable systems, as this was the first internal AI initiative. Also advised on best practices and how/where AI can be applied most effectively, while avoiding low-impact/high security risk applications.- Developed an LLM service that takes customers' natural language and returns SQL queries to provide custom results.- Created a Flask API to interact with the LLM, training interface, and custom plot generation.- Researched stacked ensembles with BERT, random forest, and XGBoost for data analysis and targeted client advice to improve customer retention.

RIMSE Fellow at UCSD's ATLAS group for computational research in materials physics. Created graph-neural-network-based software for the prediction of dynamic mechanical loss of a material. Adapated "Properties from Artificial Neural Network Architectures" (PANNA) to train a high-dimensional neural network potential for MOFs.