PORTFOLIO

WEB DEVELOPMENT

• HIS HOLINESS WILL SEE YOU NOW
  
  Are you in need of forgiveness? Has the crushing weight of original sin got ‘ya feeling blue? Well, cross yourself and step into the confession booth, HIS HOLINESS WILL SEE YOU NOW.
  
  HIS HOLINESS WILL SEE YOU NOW is an interactive audience with the Pope. Users can speak to the Pope in realtime, offering confession of the sin of their choosing OR they can just chat. The Pope is here to listen. Users uncertain as to how they may have sinned can choose from convenient, pre-written lists of VENIAL SINS and MORTAL SINS. Just don’t forget to hit the ABSOLVE button before you leave. You wouldn’t want to leave uncleansed…
  
  Featuring: React 19, TypeScript, Vite, Tailwind CSS, OpenAI API, GPT 4.1-nano, Three.js


• THE TEXAS A&M FOOTBALL HALL OF FAME
  
  Enjoy the longstanding tradition of good old fashioned traditional traditions in College Station. Take pride in feeling proud of your pride. This land is your land. Vote with your wallet.
  
  Featuring: React 19, TypeScript, Vite, Tailwind CSS, OpenAI API, DALL-E 3, GPT 4.1-nano


• WOKE OR NOT WOKE

WRITING

• THE OFFICIAL FUNDAMENTAL DATUM SUBSTACK

The only place where you can read official fundamental datum-approved literature like What Is An Altar?, a story about quantum physics, Faustian bargains, Major League Baseball, and a man whose hand won't stop exloding.

DATA SCIENCE

Pythonic data science projects built using Jupyter Notebooks.

• Natural Language Processing: IT Support Ticket System
  
  This project leverages a powerful language model, utilizing Python and Pydantic for structured data modeling, to automatically classify IT support tickets by category, urgency, and sentiment. By extracting key information and providing confidence scores, the system enables support teams to prioritize and route tickets more effectively. This streamlined approach, built with a modern tech stack, reduces response times, improves customer satisfaction, and optimizes workforce allocation for a more efficient IT support workflow.
  
  Featuring: EDA, Large Language Models, Transformer Models, Prompt Engineering


• Computer Vision Image Classifier: Cargill Agriculture
  
  This project addresses the agricultural challenge of manual plant identification by building a computer vision model to automatically classify 12 different species of plant seedlings. A Convolutional Neural Network (CNN) was developed in Python using TensorFlow and Keras, employing data augmentation and transfer learning with a pre-trained ResNet50 model to achieve robust performance. The final model demonstrates the feasibility of this approach, providing a foundation for real-world applications such as mobile-based plant identification and automated field monitoring to enhance precision agriculture.
  
  Featuring: EDA, Image Processing, Keras, Tensorflow, Convolutional Neural Networks, Transfer Learning


• Neural Networks Churn Predictor Model: Pinnacle Bank
  
  This project tackles the critical issue of customer churn by building a predictive model for Pinnacle Bank to identify clients at risk of leaving. Using TensorFlow and Keras, an Artificial Neural Network was developed, incorporating data preprocessing, SMOTE to address class imbalance, and Dropout regularization to prevent overfitting. The resulting model not only accurately predicts customer churn with a high recall rate but also provides actionable business recommendations to help the bank implement targeted retention strategies.
  
  Featuring: EDA, Data Cleaning / Preprocessing, Tensorflow, Keras, Artificial Neural Networks, Regularization


• Machine Learning Personal Loan Predictor Model: Citizen’s Bank
  
  This project develops a predictive model for a Citizens Bank marketing campaign to identify liability customers with a high probability of purchasing a personal loan. Using Python libraries such as pandas and seaborn for data cleaning and exploratory analysis, a Decision Tree Classifier was built with scikit-learn to predict loan acceptance. The final model was optimized through hyperparameter tuning and pruning, successfully identifying key predictors like income and education to create a targeted marketing strategy.
  
  Featuring: EDA, Data Cleaning / Preprocessing, Missing Value Treatments, Decision Trees, Pruning


• Exploratory Data Analysis: DoorDash
  
  This project conducts an exploratory data analysis (EDA) on DoorDash order metadata to uncover key trends in customer demand and restaurant performance. Using Python libraries such as Pandas for data manipulation and Seaborn for visualization, the analysis applies univariate and bivariate techniques to examine variables like cuisine type, order cost, and delivery times. The resulting data-driven insights culminate in actionable business recommendations for DoorDash to enhance customer experience and drive revenue.
  
  Featuring: EDA, NumPy, Pandas, Seaborn, Univariate Analysis, Bivariate Analysis