projects

Research systems spanning experimental decision-making, interpretable modelling and scientific data engineering.

These projects form a connected research programme: experimental evidence feeds probabilistic and interpretable models; those models guide the next formulation or experimental decision; literature-mining systems extend the evidence base. Code and reproducibility materials are available where data-sharing constraints allow.

Adaptive Batch Optimisation

Active learning for dermal formulations using GPR, expected improvement and hypervolume contribution to adaptively balance performance and information gain.

Early Decision-Making for IVPT

Probabilistic forecasting from partial permeation trajectories to stop unpromising experiments early while accounting for uncertainty.

Domain-Constrained Symbolic Regression

Interpretable dynamic models of dermal drug release that embed domain constraints, support optimisation and expose scientifically meaningful relationships.

SkinMiner

An LLM-powered data-mining framework for traceable extraction, verification and structuring of dermal formulation evidence from the scientific literature.

PhD Research Hub

The umbrella repository connecting publications, code, datasets and reproducibility materials across my doctoral research programme.