Research Interests

My research spans multiple disciplines, covering a wide range of topics. My primary focus is the study of deep learning architectures from mathematical perspectives. Currently, I am particularly interested in the following topics:

• Dynamical properties of pre-trained language models,

• Equivariant models for processing geometric data (graphs, meshes, point clouds, and neural networks),

•  Multimodal models for biology, robotics, and computer vision,

•  Adaptation of large language models to downstream tasks.

In addition to my primary research on deep learning models, my work also addresses computational problems in algebraic differential equations, as well as arithmetic and geometry over skew-fields.