My bachelor's thesis at the Technical University of Cluj-Napoca compared four machine-learning models for emotion classification. A fine-tuned DistilRoBERTa model achieved 92.2% accuracy, and I designed a rule-based expressive-intensity module and a Streamlit app for real-time prediction and model comparison.
As a Software Development Intern at Exigea (INFOWORKS SOLUTIONS SRL), I built React front-end pages and tested the site across browsers and screen sizes. My project work also includes a Python vocal pitch-correction application and a C client-server network application bridging IPv4 and IPv6.
