Created an LLM-based pipeline to analyze political sentiment in large text corpora as part of the Digital Markets Initiative's research into the economic and political impact of foundational models. Used cosine similarity on embeddings generated via API calls to the UF HiPerGator supercomputer.

• Built an LLM inference pipeline using API calls to UF's HiPerGator supercomputer for large-scale text processing
• Implemented cosine similarity analysis on sentence embeddings to cluster and classify political sentiment
• Applied Scikit-Learn, NumPy, and TensorFlow for model training and embedding generation
• Research contributing to ongoing work on the economic impact of foundational AI models