Home
My Background
I’m an AI/ML researcher in the field of cheminformatics and computational modeling for drug discovery. I began my journey with a Master’s degree in Pharmaceutical Chemistry, followed by invaluable experience at AstraZeneca, where I worked across computational chemistry and bioinformatics. Driven by a fascination with generative methods, I pursued a PhD at the University of Cambridge, focusing on de novo molecule design using machine learning, under the supervision of Prof. Andreas Bender and in collaboration with Nxera Pharma. I now continue this work as a postdoctoral researcher at Universitat Pompeu Fabra, under the guidance of Prof. Gianni de Fabritiis and in collaboration with Johnson & Johnson Innovation.
Research Interests
I’m driven by the challenge of how we can safely and efficiently make targeted modifications to the incredibly complex systems that constitute human biology. I believe that leveraging AI and machine learning in this endeavor offers our best chance of making meaningful progress. In the long term, I’m interested in any and all technological advancements that help move us closer to that goal — though I know achieving it will require the collective effort of many.
In the short term, I aim to contribute by improving how we explore chemical space to identify small molecules with desirable biological effects. More specifically, I’m interested in:
Developing generative machine learning to efficiently search expansive chemical spaces
Understanding how to ensure that generative models have sufficiently explored their implicit chemical space — in other words, how can we be confident we’ve identified all molecules of interest?
Integrating protein structure-based methods to improve the binding affinity of de novo designed molecules
Setting standards for the evaluation of generative machine learning models
Recent Highlights
Featured Publications
- Test-Time Training Scaling for Chemical Exploration in Drug Design
- Modern hit-finding with structure-guided de novo design: identification of novel nanomolar adenosine A2A receptor ligands using reinforcement learning
- MolScore: a scoring, evaluation and benchmarking framework for generative models in de novo drug design