Tatiana Gaintseva

I am a PhD candidate in AI at Queen Mary University of London, supervised by Prof. Greg Slabaugh and Dr. Martin Benning, and supported by a DeepMind studentship. In 2024–2025, I was a Research Intern at Huawei Noah’s Ark Lab in London, working with Ismail Elezi and Jiankang Deng on steering and control of generative models.

My research focuses on understanding intermediate representations in foundation models and developing principled methods for controlling model behavior. I am especially interested in representation geometry, interpretability, activation steering, and concept-level interventions in large generative models.

Previously, I obtained my BSc and MSc degrees in Applied Mathematics and Data Analysis from the Moscow Institute of Physics and Technology, and completed the master’s-level programme at the Yandex School of Data Analysis. My background combines machine learning, mathematics, and generative modeling.

Alongside research, I have extensive experience in teaching fundamentals of AI, as well as building educational programmes at scale. I am a founding member, methodologist, and lecturer of Deep Learning School at MIPT, and have taught ML/DL courses at Moscow State University, MIPT, and Yandex School of Data Analysis. I currently teach AI and LLM fundamentals at Nebius Academy in London.