My research interests lie in the application of computational learning algorithms to scientific data. I have applied this technique in many different areas of science. During my undergraduate degree at The University of Bath, I performed optimisation of the electrostatic field on a crystal surface to obtain electron probability densities. The technique involved iterative line optimisation using parabolic fitting, conjugate gradient techniques and preconditioning.

During my PhD at University College London, I worked on the Functional Oxide Discovery project developing algorithms for the prediction of electroceramic material properties including permittivity and oxygen diffusion. Artificial neural networks were successfully used to obtain property predictions from compositional data. Furthermore, genetic algorithms were applied to 'invert' the neural network, resulting in compositional predictions of desirable materials. I also developed a database of ceramic materials containing their composition, processing parameters and properties which forms a useful resource for the scientific community

My work as a research engineer at MIT includes the development and maintenance of the MIMIC-II database.