Building at the intersection of systems, data & AI

Pursuing a B.S. in Software Engineering and Data Science at Loyola University Chicago — GPA 3.1, two semesters on the Dean’s List. Graduated from Novi High School with a 4.0 weighted GPA, 1400 SAT, President’s Award for Educational Excellence, AP Capstone Diploma, and College Board Hispanic National Recognition.

Completed the Google Data Analytics Certificate in one week. Earned the Football Data Analytics certification at Universidad Europea de Madrid Escuela Universitaria Real Madrid, applying ML and visualization tools to KPI development in sports contexts.

Fluent in English and Spanish, conversational in French and Italian. Two years volunteering as an acting coach and assistant director. Tutoring experience in English and advanced mathematics at high school and middle school levels in Chicago.

As a Software Engineering Intern at BCS Automotive Interface Solutions, I built a haptic feedback device testing GUI using Python and Tkinter — a platform for engineers to design, visualize, and fine-tune vibration waveforms for automotive control systems.

The application supports real-time waveform customization via adjustable sliders and Matplotlib-powered plotting. Users can generate and manipulate response curves, save them to CSV files, and load preset waveforms for reuse or calibration. Supports ERM, LRA, piezoelectric, and electromagnetic actuator types.

Serial port integration enables live interaction between software and hardware. The UI enables or disables actuator tabs based on COM port connectivity, with modes for basic, advanced, and audio-to-vibe configurations.

The embedded backend was initially built in Arduino, then ported to C++, handling the control logic for interpreting incoming serial commands and translating them into actuator-level responses.

Built as part of NextGen Advanced AI, a startup co-founded with fellow Loyola students. I fine-tuned DeepSeek-R1-Distill-Qwen-1.5B from Hugging Face using LoRA (Low-Rank Adaptation) — updating the model with minimal compute by injecting lightweight, trainable adapters into a frozen base model.

Optimized for the MPS (Metal Performance Shaders) backend on macOS. The training script loads prompt-completion pairs from a JSON file, formats them for instruction tuning, and saves the LoRA adapter. Fine-tuning focused on philosophical reasoning.

Deployment uses a Flask API serving the model through a /generate endpoint with streamed responses. Frontend rebuilt with React and Vite, using Axios for communication and Docker for environment consistency.

Futstat — a full-stack Premier League analytics platform built for COMP 330, Software Engineering, at Loyola. As Project Lead, I coordinated a team of five, set the architecture, and managed version control across frontend, backend, and data science components.

Built with React.js, Node.js / Express, and PostgreSQL, integrating real-time football data feeds. I wrote Python ETL pipelines to populate the system with 25 years of historical match data.

The ML system uses scikit-learn with logistic regression, random forests, and XGBoost, trained with temporal cross-validation to prevent data leakage. Models are serialized via pickle and integrated into Express through Python child processes for real-time RESTful inference.

Project lead and primary architect for a web-based major recommender system built in collaboration with Dr. Catherine Putonti at Loyola University Chicago — a consulting project integrated into Loyola’s existing tech stack.

I designed the overall project structure and the core algorithm to match student coursework against scraped major requirements, generating personalized academic recommendations. I built the data pipeline, recommendation logic, and Flask-based UI.

Led the team throughout — technical communication, task coordination, and keeping delivery aligned with the advisor’s vision.

An internal tool organizer and database website built during my internship at BCS Automotive Interface Solutions using PHP and MySQL, to manage internal workflows, member profiles, and content dynamically. Hosted locally via XAMPP for rapid prototyping.

Responsive frontend with HTML/CSS/JavaScript; backend handles user authentication, session management, and data-driven interactions including event registrations, contact forms, and content submission portals.

Built with reusable PHP components and secure database interaction via parameterized SQL queries to prevent injection vulnerabilities. Supports file uploads, email notifications, and full CRUD operations.

A multiple regression analysis exploring the statistical factors that best predict a Premier League team’s final league ranking during the 2023–24 season, across a dataset of 75 performance metrics.

All predicted ranks fell within a reasonable range of the actuals, with no residual exceeding 3× RMSE. Notably, a higher dispossession rate correlated with better rankings — reflecting the aggressive pressing systems of elite clubs like Liverpool rather than disorganization.

A Python-based soccer match and 1v1 predictor combining real-world data analysis, machine learning, and a GUI to simulate and predict outcomes for both team matchups and individual player scenarios. Built with pandas, numpy, scikit-learn, and tkinter.

Three main components: data preprocessing, logistic regression and Monte Carlo simulation for probability estimation, and a dual-tab tkinter GUI with team matchup and 1-on-1 player predictor tabs.

Users can view win probabilities, success rates, common formations, and tactical information. Real-time player search and automatic team mapping round out the usability.