Computer-Aided Molecular Design
Most broadly, I am interested in how computation can be leveraged in the design of organic and biological molecules for biomedical and nanotechnological applications. Many of the properties that make these molecules attractive as therapeutics, catalysts, or sensors depend on the complex interplay between intra- and inter-molecular interactions, and solvent, electronic, and entropic effects. Using data-driven and physics-based approaches, I aim to predict these molecular properties, as well as to understand their atomic-level determinants.
Machine Learning and Experimental Design
I am interested in the application of machine learning algorithms to property prediction, chemical space exploration, and experiment planning. These include discriminative models like random forest, Gaussian processes, and neural networks, for predicting molecular properties; Bayesian optimization algorithms to guide efficient, possibly autonomous, experimental and computational studies; and generative models like variational autoencoders and generative adversarial networks for inverse molecular design.
Molecular Dynamics and Biophysics
I use Molecular Dynamics (MD) and Monte Carlo (MC) simulations to study at the atomic level the multitude of conformations available to molecules in solution, and to extract information about their thermodynamic, kinetic, and dynamical properties. Being rooted in physical principles, these simulations allow for the atomic-level interpretation of the role of dynamics, inter-molecular interactions, and solvation in biological function and dysfunction.