Motivation
My research aims to support better, fairer, and more informed decision-making. Much of my work addresses credit-related challenges, many of which reflect issues common to decision-making in other domains, too. The methods I use and develop are built for complex settings where uncertainty, dynamics, and prediction intersect.
Research Landscape
My work spans several interrelated themes:
Methodological Challenges
- Dynamic prediction in time-evolving settings
- Uncertainty quantification for risk-sensitive decisions
- Interpretability and explainable AI (XAI)
- Stress testing under extreme or systemic scenarios
- Spatial dependence in entities behaviour
- Selection bias from non-random data access (e.g., accepted loans only)
- Imbalanced data and rare-event learning
- Scalability and variable selection to high-volume, high-dimensional data
- From association to causation in predictive modelling
Emerging & Societal Challenges
- Climate change and biodiversity loss as systemic risk factors
- Sustainability in resource allocation and financial markets
- Fairness and equitable access to financial services
Data Modalities
- Cross-sectional observations
- Longitudinal (panel) data
- Time-to-event data in continuous or discrete time
- Unstructured data such as text, transactions, and network traces
Methodological Approaches
- Bayesian hierarchical modelling
- Distributional regression frameworks
- Survival analysis including multi-state models, cure models, and joint models for longitudinal and survival data
- State-space models
- Causal inference
I am currently accepting PhD students. If you’re interested in working on related topics, feel free to get in touch.
MSc dissertation students. If you are looking for a dissertation project aligned with the themes above, I’d be happy to discuss options.
Some available topics: view the current list here (you may need a UoE account to access it).
If you’d like to work with me, please email me with a short description of your background and your interests (1–2 paragraphs).