Duke University Researcher

Implemented a GPT model on a live server to provide limited coding feedback for student assignments. Developed an API for Duke University websites, enabling real-time feedback integration.

March Madness Model

Led a research project to identify key metrics predicting March Madness success (offensive & defensive efficiency). Developed a K-Nearest Neighbors (KNN) model to evaluate prediction accuracy. The model achieved 37.2% accuracy, far surpassing the 14.2% baseline for random guessing.

Tractor Supply Product Locator

Search for products on the Tractor Supply website, and retrieve detailed information to display in a GUI window. Web scraping using BeautifulSoup, formatting, and managing product attributes, fetching POG (Planogram) data from a CSV file, and displaying product details and related images through a user interface created with Tkinter.

Website Development

Created full-stack websites for personal use and clients. Includes functional checkout modal. Extensive HTML, CSS, SQL, Flask, Node.js, Stripe, and JavaScript to develop full-stack websites.

Neural Network

Developed and implemented a neural network framework in Python, featuring custom classes for neurons and synapses with randomly initialized weights and biases. Built in-depth understanding of NN and how they operate. Implemented PyTorch to predict, with high accuracy (2.5 points per player range), NBA player's average points per year.

Chess960

After listening to a Magnus Carelson podcast, was inspired to prove that Chess960 actually hed 960 different starting combinations. Created an algorithm to find all of these starting combinations.