Georgy studied Mathematics for his bachelor’s at University College London and then proceeded to study Data Science and Machine Learning for his master’s degree there as well. His dissertation focused on improving the efficiency of Large Language Models at inference, showcasing the improvements on Llama-2, by being able to reduce model size without a significant decrease in performance.
Georgy’s research interests lie in high-dimensional statistics, uncertainty quantification, and the theoretical foundations of machine learning. Outside academia, he enjoys combat sports such as boxing and kendo, as well as climbing and strategy games.
