CV

Work Experience

Lead Climate Data Scientist, NatWest Group, London
Oct 2025 – Present

Climate Risk Modelling Technical Lead

• Leads the technical direction of the bank’s bottom-up climate risk modelling capabilities across Corporate, Mortgage and Commercial Real Estate portfolios, covering physical and transition risks and translating business and regulatory requirements into delivery priorities.
• Provides technical oversight of model development, reviewing code and merge requests, determining modelling approaches and signing off production outputs to ensure analytical quality.
• Applies knowledge of financial climate risk frameworks and methodologies, including NGFS scenarios, RCPs, GCMs and return-period-based hazard metrics, to support robust risk assessment.
• Acts as a recognised physical climate risk SME across Retail, Corporate and Commercial Real Estate, advising stakeholders on data methodologies, limitations and interpretation; collaborate with Credit Risk to feed model outputs into downstream probability of default (PD) models.
• Leads the technical evaluation of external physical risk data providers, defining requirements, coordinating assessments and engaging with vendors to evaluate their methodologies and capabilities.
• Line manages eight colleagues in a multidisciplinary team of approximately 20 across data science, data engineering, MLOps, business analysis and delivery; lead quarterly planning and contribute to annual budget and resource planning.
• Introduced production readiness standards, automated operational checks and clearer documentation, release and sign-off controls; coordinate interdependent workstreams to support reliable model delivery and ongoing operation.

Senior Climate Data Scientist, NatWest Group, London
Mar 2023 – Oct 2025

Stream Lead of the Mortgage Climate Impact Model

• Led development of the Mortgage Climate Impact Model, quantifying acute and chronic physical and transition risks to estimate property value haircuts for downstream credit risk modelling.
• Developed methodologies to capture climate impacts over a 10-year horizon and assess the effect of Flood Re on mortgage probability of default; led design and development of the mortgage transition risk model with an external consultancy.
• Refactored the model into a modular, object-oriented architecture, improving maintainability, flexibility, documentation and automated testing.
• Applied anomaly detection and data analysis to identify input quality issues and fill data gaps, and used Bayesian optimisation to calibrate wind damage curve parameters.
• Led the Internal Model Validation submission, compared outputs against vendor average annual damage (AAD) estimates. Model outputs support climate disclosures, annual stress tests and regular reporting, with results communicated to non-technical stakeholders.

Principal Data Scientist (G7), UKHSA (formerly JBC), London
Sept 2021 – Mar 2023

Lead Data Scientist of the Epidemiological Ensemble Modelling Team

• Joint technical lead of an epidemiological ensemble modelling team of approximately five, coordinating delivery of national consensus estimates of the effective reproduction number (R), medium-term projections and scenarios under alternative R values.
• Led the technical execution of multiple epidemiological models, developing Python and R wrappers and an agent-based model, and coordinating model runs across different modelling approaches.
• Synthesised outputs from diverse models into consensus estimates, assessing differences in assumptions, data sources and model behaviour to interpret divergent results and communicate uncertainty without dismissing legitimate alternative projections.
• Presented modelling results and methodological considerations to academic stakeholders to support agreement on consensus estimates, which informed government COVID-19 briefings and were published for public use.
• Led development of a Markov-chain model to estimate the effectiveness of the NHS Test and Trace programme, producing the first model published for the Joint Biosecurity Centre (JBC).
• Improved the reliability and reproducibility of modelling workflows through standardised packaging, version control practices, Docker, GitHub Actions and automated validation. Enabled overnight, cloud-based model runs with human oversight and traceable code, data and output artefacts, reducing the need for overnight manual monitoring.
• Co-authored research on combining epidemiological models to generate a consensus effective reproduction number for the COVID-19 epidemic in England.

Data Scientist, Avanade (previously Altius), London
Nov 2019 – Sept 2021

• Developed reusable Python components for an object-oriented data science framework, improving code reuse and the efficiency of model development and productionisation.
• Built an NLP pipeline for handwritten prescription processing, achieving 98% accuracy across 60% of prescriptions, and applied text analytics, including TF-IDF and Latent Dirichlet Allocation (LDA), to analyse consultation responses and identify key themes and sentiment.
• Developed a speech-enabled restaurant ordering application using speech recognition, NLP, Flask and SQL, integrating language processing with a database-backed application.

Summer Research Project, Met Office, Exeter
Summer 2018

Developed and tested finite-volume and interpolation methods for the transport equation, investigating numerical inaccuracies arising from a new grid discretisation for the Met Office’s next-generation Unified Model. Identified sources of error and proposed potential solutions, documenting findings in a technical report.

Education

PhD in Applied Mathematics, Imperial College London, Mathematics of Planet Earth CDT
2016 – 2020

Thesis: Lagrangian Transport in an Idealised Meandering Oceanic Jet

Investigated methods for parameterising unresolved small-scale particle transport in an idealised ocean jet, motivated by the role of turbulent processes in heat transport across major ocean currents. Evaluated fractional Brownian motion and dimensionality reduction of chaotic ocean dynamics, identifying limitations in their ability to represent the underlying transport processes. Developed and validated a more accurate numerical method exploiting properties of the governing equations, using analytical solutions to assess accuracy. Ran a complex quasi-geostrophic ocean model in Fortran on high-performance computing clusters and collaborated on modularising the codebase to improve maintainability and usability for other researchers.

Masters in Research, Mathematics (Distinction), Imperial College London, Mathematics of Planet Earth CDT
2015 – 2016

Courses included: Data and Uncertainty, Geophysical Fluid Dynamics, Computational Stochastic Processes, Numerical Methods, Dynamical Systems and Partial Differential Equations.

MSci Mathematics (1st Class), University College London, Department of Mathematics
2011 – 2015

Thesis: Vortex dynamics in a uniform flow around two cylinders