Rocío Mercado Oropeza | Assistant Professor at Chalmers University of Technology

Generative and predictive machine learning models are reshaping how we explore chemical compound space, yet meaningful impact requires models that can reason across molecular scales, handle heterogeneous data, and adapt to real-world design constraints. In this talk, I will present recent work from our group on AI-driven molecular design, spanning small-molecule therapeutics, targeted protein degraders, and sustainable materials for emerging technologies. A central focus of this talk will be on synthesis and biological fate: I will discuss RetroSynFormer, a Decision Transformer for multi-step retrosynthesis planning, as well as LAGOM, a transformer-based chemical language model for predicting drug metabolites. The models we develop illustrate how AI can support chemists not just in designing molecules, but in planning how to make them and anticipating what happens to them in a biological system. More broadly, we are developing approaches that combine generative modeling, structure-aware learning, and chemically informed representations to support molecular design tasks across various length scales and domains. Our work is carried out in close collaboration with industrial partners across pharmaceuticals and advanced materials (including AstraZeneca, Intel, and Merck), ensuring that the methods we develop translate to practical design workflows.

--

Please complete the registration form and you will receive a link after confirmation.