Academic work · Research · Innovation
Research, university projects and innovation work presented through the question, my contribution, the method and the available evidence.
Independent and collaborative investigations grounded in experimentation, evaluation and reproducible technical work.
Improving childhood poverty prediction by combining satellite imagery, geographic encoding and systematic data quality assessment.
Key evidenceThe published pipeline reduced MAE from 0.2167 to 0.1759, an 18.83% relative improvement, and achieved 0.1658 MAE across 33 African countries.
Designing and evaluating retrieval-augmented generation pipelines to investigate bias mitigation in large language model systems.
Key evidenceBuilt automated experiment tracking for RAG evaluations using open-source datasets and cloud GPU infrastructure.
Substantial technical work completed through university study and assessed research.
Developing diffusion models as plug-and-play priors for image restoration and inverse problems.
Key evidenceImplemented a diffusion model within a plug-and-play framework using a variable splitting method.
Fast-moving projects that combine evidence, commercial thinking and clear recommendations.
Exploring what London can learn from global rail-led development models to deliver more housing around transport infrastructure.
Key evidenceTranslated international evidence into six London proposals covering land readjustment, over-track development, tradeable density rights, rail co-investment, public ground leases and car-free station areas.
Shaping the commercial strategy for an AI-enabled healthcare mobility product.
Key evidenceTranslated technical capabilities into a clear commercial value proposition and go-to-market recommendation.