Data scientist and machine learning engineer with a PhD in fluid dynamics and 7+ years of research experience. I build agentic AI and LLM systems for scientific work, and apply Bayesian methods and deep learning to physical and engineering data, with tested, reproducible code.
- AI Literature Review Assistant: an agentic LLM system (PydanticAI) that turns research PDFs into citation-grounded literature reviews, with sentence-level traceability. Live demo.
- Boiling-point prediction, active learning & Bayesian optimisation: nested cross-validation, a label audit, and BoTorch Bayesian optimisation that finds 69% of in-spec compounds after 12% of experiments.
- Two-stage movie recommender: TFX two-tower retrieval and LightGBM LambdaRank re-ranking, served by FastAPI in ~70 ms per request. Live demo.
- Vehicle damage detection: YOLO11 with MLflow tracking, a containerised FastAPI service, and CI with champion/challenger model promotion.
PhD in Fluid Dynamics & CFD, Cranfield University. Nine peer-reviewed journal articles, eight as first author, in fluid dynamics, multiphase flow and interfacial science. Google Scholar