Research · Website · Software · ORCID · LinkedIn
I am a PhD researcher at Imperial College London working at the intersection of energy systems, machine learning and electricity markets.
My research develops computational methods for spatial energy planning, power-system emissions and low-carbon infrastructure. I am particularly interested in using graph learning, geospatial analytics and statistical learning to add useful spatial and temporal information to energy-system models while keeping the underlying optimisation and engineering logic interpretable.
I also work with Climate Compatible Growth on energy-system research supporting Zambia and Malawi.
|
Settlement-scale graph-based aggregation for electrification planning in Zambia. A variational graph autoencoder learns spatial representations from a 24,550-settlement proximity graph. The resulting embeddings are converted into connected, technology-characterised planning units and interpretable supply-curve bands. Python · PyTorch Geometric · VGAE · Geospatial ML · Energy Planning |
Research on spatial marginal emissions in transmission-constrained, decarbonising power systems. The work combines power-system modelling and interpretable machine learning to understand how network constraints, dispatch conditions and spatial demand changes shape marginal emissions signals. Machine Learning · Electricity Markets · Power Systems · Emissions |
| Energy-system planning | AI and machine learning for energy | Electricity markets and emissions |
|---|---|---|
| Capacity and infrastructure planning | Graph representation learning | Marginal emissions |
| Spatial electrification | Geospatial machine learning | Transmission constraints |
| Open-source energy models | Statistical learning | Market and system signals |
| Low-carbon transitions | Interpretable AI | Decarbonising power systems |
Python · PyTorch · PyTorch Geometric · scikit-learn · optimisation · graph neural networks · geospatial analytics · time-series analysis · PyPSA · OSeMOSYS · OnSSET · GeoPandas · NetworkX
I aim to make research outputs reproducible and reusable. My public research repositories are structured around versioned inputs, transparent assumptions, deterministic analysis where possible, validation checks, tests and citation metadata.
The Zambia VGAE release includes frozen analysis artifacts, checksum verification, reproducibility scripts, tests, source attribution, data licensing and a CITATION.cff file.
- PhD Researcher, Imperial College London
- Research Assistant, Climate Compatible Growth
- Energy-system research supporting country partnerships in Zambia and Malawi