YURIY PEREZHOHIN is a PhD. student of Data Science at NOVA IMS, Lisbon, Portugal, where he is developing new methodologies for reducing the computational resources of Large Language Models in their training or inference process. His research interests include Natural Language Processing, Code Generation, and cross-modal embeddings. He has developed during his Master’s Thesis an innovative approach to tackle text-to- SQL task, where the model is free of training and adaptable to new domains. He also lectures practical classes on deep learning at NOVA IMS.