Vera aims to develop theoretically grounded machine learning algorithms for biology, chemistry, and related engineering fields, hypothesizing that these technologies can enable more efficient and controllable solutions.
Her master's thesis explores long-range sequence models, investigating their mechanisms and architectural designs. The goal is to create models that can effectively generalize across the fundamental modalities of the central dogma in molecular biology. Additionally, Vera is also interested in how these models might provide mechanistic insights that could enhance domain scientists' understanding of complex systems.