Best Romanian AI Thesis - BRAIT 2026

UTCN and the Computer Science Department are proud to see the work of Antonio Alexoaie is recognized at the national level. “Hyperflux: Pruning Reveals the Importance of Weights” has been named Best Romanian AI Thesis at BRAIT 2026 and was presented at Romanian AI Days. The thesis explores how pruning can reveal the importance of weights in neural networks, offering an interesting perspective on understanding the internal structure and behavior of AI models. My congratulations also go to the Romanian AI Days / BRAIT 2026 organizing team for creating a platform that recognizes and promotes the next generation of Romanian AI researchers: Adriana Stan (UTCN), Alexandru Sorici (UNSTPB & ARIA), Elena Burceanu (Bitdefender & UNSTPB), Emanuela Haller (UNSTPB), Florin Brad (Bitdefender), Sorin Grigorescu (RovisLab & Transilvania University), Razvan Pascanu(EEML & DeepMind), Traian Rebedea (NVIDIA & UNSTPB), and Viorica Patraucean Pătrăucean (EEML & DeepMind). It is encouraging to see researchers from universities, industry, startups and leading international AI organizations coming together around the development of Romania's AI research community. Congratulations to the students! 🎓👏 I look forward to seeing where his research journey takes them next, as Antonio is now a master student at Technical University of Munich.