BACKGROUND

Two postdoctoral positions, each for 3 years, are available through UKRI funding underwriting a European Research Council (ERC) Advanced Grant. This position is open to outstanding and ambitious computational scientists to work in molecular simulation progressing the in-house machine-learnt potential called FFLUX. One post (A) will focus on considerably speeding up the in-house code DL_FFLUX, written in FORTRAN90, and will carry out the simulations. The second post (B) will advance the machine learning behind the FFLUX potentials. By its novel architecture, the locally developed force field FFLUX aims to make a step change in the reliability of modelling of peptides/proteins in aqueous solution. The machine learning method Gaussian Process Regression is used to create knowledgeable quantum atoms that accurately predict atomic energies and multipole moments. It is vital that the two naturally complementary postdocs collaborate.

The overall task is to advance the accuracy and performance of the in-house machine learning force field FFLUX.