✦ stars over time

I work on how coherent behaviour comes out of local interactions, and on how to make causal claims about it.

I came to this from economics. What I liked there — simple models that explain a lot, and the moment an identification strategy lets you actually claim something — turned out to be what I like in complex systems and computational biology too.

MSc in Economic Analysis: econometrics, causal inference, instrumental variables on observational data

MSc in Data Science and Machine Learning

Distributed sensing from local rules (in progress)

Can a tissue where each cell only senses its neighbours tell a one-off wound from a moving herbivore? The same cells are damaged either all at once or one per step, and the tissue's defence is scored against both possible futures. It only counts as discrimination if it pays off in both. I test minimal local rules (pulse counting, priming, novelty detection), then noise, communication knockout and network rewiring. Python · NumPy · networkx

Selection vs induction in triple-negative breast cancer (planned)

After chemotherapy, how much of the surviving cell-state composition is selection of states that were already there, and how much is induced by cell–cell communication? The plan: communication programmes from longitudinal single-cell data, network analysis of their architecture, and an agent-based model with switchable communication, calibrated by simulation-based inference.

Heterogeneous treatment effects in TNBC (exploratory)

CATE methods for neoadjuvant chemotherapy response.

Thinking in Models: a reading list on modelling as a way of thinking, across economics, social science and biology

Emergence · agent-based modelling · political economics · causal inference and identification · simulation-based inference · systems biology · cancer