Computer Science student at UTSA · Break Through Tech AI Fellow
I am a computer science student with interests in machine learning, natural language processing, ethical AI, software engineering, and UI/UX design. Much of my recent work centres on the computational analysis of classical Arabic poetry, where I examine how linguistic structure can be modelled using modern machine learning techniques.
Broadly speaking, I am drawn to projects that instersect language, culture, design, and real-world data.
Computational Analysis of Classical Arabic Poetry
An ongoing machine learning project analysing sentiment, named entity recognition, and thematic structure in classical Arabic texts. The work involves building NLP preprocessing pipelines, experimenting with feature representations, and evaluating supervised and unsupervised models, with plans for a full-stack demonstration.
Chrono - Productivity Timer & Task Tracker GitHub Repository
Chrono is a JavaFX application designed to help users manage tasks, run countdown timers, and customise themes. Developed in IntelliJ using JavaFX, SceneBuilder, and Maven for dependency management, the app emphasises functionality and user-friendly interface design.
I contributed to UI design, prototyping, implementation, and testing. The project includes unit testing with JUnit 5 and supports theme switching and persistent task storage.
Prototype & design assets: Figma File
Technologies: Java, JavaFX, SceneBuilder, Maven, JUnit, FXML
ALERG’Z -- Allergy-Safe Food Recommendation App
GitHub Repository
A mobile application developed during RowdyHacks 2025, a 24-hour hackathon, where the project was awarded Best Beginner Project. The app allows users to scan food products and quickly determine allergen safety, with AI-generated alternatives suggested when products are unsuitable.
I worked on the FastAPI backend, integrating OpenFoodFacts and large language models via OpenRouter to support real-time barcode scanning, allergen detection, and personalised recommendations.
Technologies: FastAPI, React Native (Expo), SQLite, SQLAlchemy, external REST APIs, asynchronous processing.
Planned Parenthood AI Studio -- Roo Response Categorisation
GitHub Repository
A machine learning challenge project conducted with Planned Parenthood through the Break Through Tech AI Studio programme. The goal was to classify chatbot responses in order to improve escalation accuracy and reduce misinformation in sexual and reproductive health conversations.
Working as part of a small team, I contributed to data preprocessing, feature engineering, and model development using TF-IDF representations alongside structural and sentiment-based features. Several models were evaluated, with a linear SVM achieving the strongest results (76% accuracy, weighted F1-score of 0.76), exceeding the project’s baseline success criteria.
Technologies: Python, Pandas, scikit-learn, Docker, Jupyter, NLP pipelines.
Languages: Python, JavaScript, C/C++, Java
Machine Learning & Data: Pandas, NumPy, scikit-learn, Matplotlib
Web & Mobile: React, React Native (Expo)
Tools: Git, Docker, SQL
- Reading & staying informed - especially on technology, culture, and world affairs.
- Biking & staying active - a healthy way to refresh and maintain focus.
- Exploring my creativity - from UI/UX design to music and aesthetics.
- LinkedIn: https://www.linkedin.com/in/fabianfigueroajr/
- Email: [email protected]
