Built a platform with code-free tuning and personalization of AI agents for dental offices to automate their phone calls and workflows.

Held office hours and graded assignments for COS 484: Natural Language Processing, COS 324: Introduction to Machine Learning, and COS 226: Algorithms and Data Structures

Assisted in developing Lang2LTL, a novel large language model framework, to translate English navigation instructions to Linear Temporal Logic (LTL). Achieves an accuracy of 81.83% on this task, outperforming the previous SOTA.

Researched Convolutional Neural Networks, Residual Neural Networks, and Deep Convolutional Generative Adversarial Networks

Developed a Flask audio-tagging application that collects songs from the Spotify API and stores crowdsourced labeled data describing songs in PostgreSQL databases.Implemented an Express app that communicates with Nautilus Software's Nemobot chatbot to query users for their preferred songs and to build a personal taste vector which is a numerical representation of the users' tastes.Used the K-Nearest Neighbor machine learning algorithm with the crowdsourced song data to provide users with a curated playlist based on their taste vectors.