At QU School of Computing & Engineering, I designed, implemented, and evaluated reproducible Python machine-learning pipelines for a peer-reviewed paper. Benchmarking classification algorithms with cross-validation, I improved F1 score by up to 29% and specificity by 22% over baseline classifiers through hyperparameter tuning.
For my senior capstone with Óbuda University, I’m building an AI system that detects and localizes people in thermal imagery for fire emergency response, with results sent to a mobile web dashboard. I’ve also built a full-stack reservation app with React, TypeScript, Express, and MongoDB, and developed an Android fitness app in Kotlin.
