PhD student, website.

Non-linear General Relativity for Stochastic Inflation, supervised by Paul Shellard and Gerasimos Rigopoulos.

This PhD project will advance the early universe theory of inflation, which explains the origin of galaxies and large-scale structure in the Universe, by developing stochastic methods with nonlinear evolution to make quantitative statistical predictions from inflationary models. Non-Gaussian signatures are inherently present in primordial fluctuations due to the non-linearities produced by both general relativity and the fact that the inflaton potential responsible for driving inflation may be interacting. The goal of this project is to analytically and numerically model nonlinear inflationary dynamics in order to make bispectrum and higher-order correlator predictions which encompass a much broader range of inflationary scenarios, including those with multiple fields. This is an important discriminant between competing inflation models that can be used to distinguish them observationally. An important element of the project will be to exploit the new GRChombo numerical relativity pipeline to model these nonlinear effects using full 3D general relativistic simulations, comparing the results with perturbative analytic expectations from effective field theory. In the last years, consequent insights have been made by the DAMTP cosmology group lead by Dr. Paul Shellard, now able to constrain a wide range of these models by using innovative separable techniques to extract and reconstruct the bispectrum (three-point correlator) from Planck satellite maps of the cosmic microwave background (CMB). Direct confrontation with new observational data sets, from both CMB experiments (Simons Observatory) and galaxy surveys (SDSS, DES etc), will guide and constrain the understanding of inflation in fundamental physics. These stochastic nonlinear methods have other wide-ranging applications, including the creation of primordial black holes and the generation of gravitational waves in the early universe.