I work on continuous and discrete optimisation problems. My PhD work focuses on the optimal control of nonlocal differential equations with applications in imaging and epidemiology. I also work on applications of mixed integer programming, like vehicle problems, healthcare, and graph partitioning.

My research focuses on Multilevel Monte Carlo Methods for Uncertainty Quantification. Specifically, I work on PDEs with random coefficients, where we aim to quantify how the uncertainty in the input propagates through the model and affects the output. This problem arises in modelling groundwater pollution, which can happen either through accidental spills or leakages from underground carbon storage or nuclear fuel repositories.

My research focuses on low-photon imaging applications by using the Bayesian paradigm. In imaging, low-photon regimes introduce severe noise processes and lead to data with high intrinsic uncertainty. To deal with these challenges, I am using MCMC methods based on Bayesian modelling.

My research is motivated by the ways insurance claims are estimated. Suppose claims amounts are independently distributed and follow a Pareto distribution, with the total number of claims following a Poisson distribution. Then the total claim amount is a compound distribution and can be easily estimated. I aim to explore the case where we remove the assumption of independence by the way of copulas and other computational methods for claims estimation.

I am modelling (industrial) processes, which can be described by particles that are submerged in a fluid, such as in nano-filtration, brewing or printing. Moreover, I am looking at the optimization of these models using optimal control techniques and implement these computationally.

I am working on new data-driven computational methods for Bayesian inverse problems, motivated by problems in image processing. These methods allow us to exploit the performance of machine learning while retaining the flexibility and interpretability of rigorous statistical models.

My research focuses on low-photon imaging applications by using the Bayesian paradigm. In imaging, low-photon regimes introduce severe noise processes and lead to data with high intrinsic uncertainty. To deal with these challenges, I am using MCMC methods based on Bayesian modelling.

I work on continuous and discrete optimisation problems. My PhD focuses on the optimal control of nonlocal partial differential equations with applications in imaging and epidemiology. I also work on applications of mixed integer programming, like vehicle problems and healthcare.

My research focuses on the theory and applications of dynamical systems that feature multiple timescales, such as the Koper model from chemical kinetics and the Hodgkin-Huxley equations from mathematical neuroscience. In parallel, I also look into the existence and stability of travelling waves in reaction-diffusion equations with cut-off functions.

My research is related to efficient statistical forecasting using Deep-Learning Reduced Order Models. In particular, I employ reinforcement learning algorithms to solve optimal well control problem in reservoir engineering.

I am modelling (industrial) processes, which can be described by particles that are submerged in a fluid, such as in nano-filtration, brewing or printing. Moreover, I am looking at the optimization of these models using optimal control techniques and implement these computationally.

My mathematical research is motivated by the need to handle energetic materials (materials with large stores of chemical energy) with care. I develop models to describe a sample's thermo-mechanical response to a given low-energy mechanical insult, focusing on the physics underlying the material's behaviour, so as to inform safety principles. When analysing the models, I tend to opt for pen-and-paper techniques, such as matched asymptotics and PDE methods.

I am working on new data-driven computational methods for Bayesian inverse problems, motivated by problems in image processing. These allow us to exploit the performance of machine learning while retaining the flexibility and interpretability of rigorous statistical models.