About Me

I am a 3rd year PhD student advised by Professor Philippe Schwaller at École Polytechnique Fédérale de Lausanne (EPFL), developing molecular generative models for accelerated drug discovery. Previously, I interned at Microsoft AI4Science in Amsterdam. Before that, I was a cheminformatics researcher in the Molecular AI team with Professor Ola Engkvist at AstraZeneca where I developed generative models for molecular design. My undergraduate degree was in chemistry, inspiring my research to harness domain expertise to augment machine learning methods for chemical discovery.

News

  • Oct. 15, 2024 - Check out another extension of Saturn showing that goal-directed molecular generation can directly optimize for constrained synthesizability using retrosynthesis models! Our method can perform multi-parameter optimization while generating molecules with valid retrosynthesis routes with the presence of enforced building blocks! Pre-print

  • July 18, 2024 - Check out an extension of Saturn showing that goal-directed molecular generation can directly optimize for synthesizability using retrosynthesis models! Pre-print

  • June 18, 2024 - Our review on generative molecular design for small molecule drug discovery is out in Nature Machine Intelligence! It is a particularly exciting time for the field as experimental validation is gradually becoming more common!

  • May 28, 2024 - We release Saturn, which is a framework for sample-efficient molecular generative design using the Mamba architecture! Check out the pre-print here. With Saturn, our aspiration is to unlock the ability to directly optimize high-fidelity oracles!

  • April 10, 2024 - Augmented Memory has been published in JACS Au! Check out a blog post about the paper here. Augmented Memory can achieve sample-efficient molecular generation optimized for docking and quantum-mechanical properties!

  • Feb. 8, 2024 - Our paper on combining active learning with reinforcement learning to improve sample efficiency in molecular generative models is out in Chemical Science.

  • Jan. 16, 2024 - Our paper on Beam Enumeration was accepted to ICLR 2024 and previously also NeurIPS 2023 AI4Science Workshop!

  • Feb. 22, 2023 - Our review paper on Bayesian optimization for chemical reactions is out in CHIMIA.

  • Feb. 4, 2023 - Our paper on Link-INVENT for generative chemical linker design is out in Digital Discovery. Check out the paper here and a Jupyter notebook example here!

  • Nov. 1, 2022 - Our paper on Icolos: workflow manager to automate computational chemistry calculations out in Bioinformatics.

  • Oct. 3, 2022 - I wrote a Jupyter notebook walkthrough for REINVENT (SMILES-based molecular generative model) as part of an awesome deep learning for chemistry educational repository by Professor Rocío Mercado. Check out the tutorial here!

  • Sep. 1, 2022 - Started my PhD with Philippe Schwaller at EPFL!

  • June 22, 2022 - Our paper on applying curriculum learning for generative molecular design is out in Nature Machine Intelligence. Check out the Jupyter notebook example here!

  • Nov. 17, 2021 - Our paper on DockStream: optimizing molecular docking scores in generative models out in Journal of Cheminformatics.