Biography
I hold a PhD in engineering and subsequently undertook postdoctoral training in computer science, a combination that underpins my current interdisciplinary work.
I am now a postdoctoral researcher in the Li Group at the Earlham Institute, where my research focuses on LLM-driven automated algorithm design.
I develop systems in which large language models retrieve relevant domain knowledge, propose candidate methods, evaluate them empirically, and distil the resulting insights back into the search process, forming a closed evaluation-driven loop that improves designs over successive iterations. To date I have applied this framework to the automated design of optimisation algorithms and physics-informed neural networks.
At the Earlham Institute I am extending this approach to the automated construction of gene prediction models. A longer-term ambition is to close the loop with the physical world. Once such systems engage directly with biological questions, each candidate evaluation becomes a real wet-lab experiment rather than a simulation.
My aim is to build a fully agentic discovery pipeline in which an AI system designs, executes, and learns from its own experiments, including through robotic manipulation of laboratory equipment. This is the central goal of my current career stage.