These days I spend my time developing and designing new models and datasets for AI drug discovery for a private company. Unfortunately this means I can't talk about my research as much as I used to.
I mostly work with structural data, developing new ways to serialise it for pretraining.
Developed a genetic algorithm, coupled with a molecular dynamics model for desiging nanoparticles optimised for cell entry. Identified rules which govern successful membrane crossing, and the dynamics associated with them.
Contributed by developing a machine learning model for automatically labelling pore in AFM images of E-coli.
Expanded upon work by Hafner et. al. by developing a new system for characterising collagen mimetic fibrils in simulation, and using evolutionary algorithms to drive their assembly.
This work made up the second half of my thesis and is currently being expanded upon for a future publication.
Development of novel experiments for investigation into immune recovery in space.
Modelling of transmission of sunlight in extraterrestrial environments.
How can we maintain healthy immune response in future Martian settlements?
Network analysis of the BCI rainforest plot to detect communities.
Analysis of the traits of subcommunities.
Are traits randomly distributed or is there some pattern?
Developed a genetic algorithm to maximise cell membrane deformation by nanoparticles in simulations.
Are there fundamental similarities between good designs? Do recognisable functional features appear?