At the University of Puerto Rico Mayagüez’s Center for Applied Ocean Science and Engineering, I built the automated data pipeline powering AlertaHAB, an early-warning system that monitors and forecasts harmful algae blooms in San José Lagoon. I developed forecasting and classification models with one- to three-week prediction horizons, achieving 0.90 recall and 0.83 ROC–AUC for one-week high-risk event detection.
At AHT Global, I contributed to a Python API, maintained MySQL queries, and helped containerize the application with Docker for deployment on AWS EC2. My work also includes satellite-based land-use analysis and a power-system stability thesis using machine learning.
