Investigated the question: What can LLM output tell us about the data it was trained on?Used various data reduction techniques (PCA, t-SNE, etc.) and unsupervised learning algorithms (KMeans, Spectral Clustering, etc.) to group LLM output into meaningful clusters.Utilized Pandas to clean and visualize large amounts of output.

Wrote complex coding problems on data analysis and visualization in the Pandas library, aimed at being difficult for LLM's to solve. Analyzed LLM output on various data analysis and visualization problems to find errors in the techniques employed.Wrote a variety of other coding problems for LLM's to train on, including topics ranging from querying using SQL, to graph algorithm problems and other applied mathematics topics.

Research assistant for Dr. Tyler Jarvis

Research assistant for Christian vom Lehn

Research assistant for Dr. Karl Diether

Business Finance (Fall 2020, Fall 2021)