Designed and trained deep learning frameworks on data from the internally reflected Cherenkov light detector (DIRC) at the Jefferson Lab. Attended weekly meetings with professors and PhD researchers to present and discuss findings. Managed code and ran jobs on William & Mary's HPC Systems.

Developed statistical model to identify demand patterns for supplies in hospital operating rooms. Backtested model on historical data and researched error metrics. Tested time series methods including ARIMA, GARCH, and Holt Winters. Built a Python library to automate Excel file upload to SQL Server using Python, SQL, and Azure Data Studio. Outlined best practices for idempotent pipelines with chief Data Engineer. Created reports for teams and executive directors throughout the health system using SQL and Python.

Led a team of four interns in updating and preparing more than 8000 iPads for the upcoming 2023-2024 school year. Used Microsoft Excel to keep track of how many iPads were updated each day, allowing for optimization of the updating process. Collaborated with supervisors and team of 20 interns to unload, process, and take inventory of more than 5000 boxes of iPads, laptops, MacBooks, Smart Boards, and monitors. Also worked alongside technicians to image more than 100 MacBooks.

Worked on applying Graph Neural Networks to reconstruct particle tracks. Helped engineer early stopping and learning rate decay features.

Completed a data science project for a Biotech startup company.- Created scikit-learn machine learning model to label medical images based on crowdsourced opinions - Used aggregating statistics and Python visualizations to compare crowdsourced opinions to expert opinions and ascertain a level of users needed for accurate data- Data-backed conclusion to show that when it comes to medical image data, crowd opinion agrees with expert opinion more often than experts agree with each other.