BSc Data Science, London School of Economics · Penultimate-year
I build systems that turn messy data into structured decisions — fraud detection classifiers, geospatial data pipelines, multi-agent AI platforms, and statistical models.
🥇 1st Place — Encode Club AI London Hackathon 2026
| Project | What it does | Stack |
|---|---|---|
| Credit Card Fraud Analysis | Analysed 284K+ transactions to surface fraud risk patterns. Logistic regression classifier with 94% recall on a 99.8% imbalanced dataset, handled via SMOTE. | Python, Pandas, Scikit-learn, SMOTE |
| ChemTrace | Multi-agent system automating synthesis route generation, real-time supplier pricing, and regulatory screening across jurisdictions. 🥇 Encode Club AI London 2026. | Python, REST APIs, ML |
| POI Activity Index | Geospatial data pipeline computing a standardised commercial activity index across 50+ London locations from OpenStreetMap, ratings, and event data. | Python, JavaScript, OpenStreetMap API |
Languages: Python (Pandas, NumPy, Scikit-learn), SQL, R, JavaScript Data & ML: Machine learning, ETL pipelines, feature engineering, data modelling, Power BI Tools: Git, Jupyter, REST APIs, FastAPI