Projects
A selection of projects and research threads that reflect my current technical profile.
Healthcare Analytics: Patient Outcome Modelling (MIMIC-III)
Large-scale clinical data processing and modelling project built on ICU patient data. Features an interactive Shiny dashboard for exploring pathogen alerts, ward heatmaps, and severity breakdowns.
- Queried, cleaned, and processed 40,000+ patient records using SQL and Python.
- Worked with missing values and irregular time-series structure.
- Built a pathogen alert classification system for clinical safety use cases.
Predicting PMSM Rotor Temperature and Torque
Machine learning project on permanent-magnet synchronous motor performance.
- Developed CNN and Random Forest models in Python.
- Engineered time-series features for thermal and torque prediction.
- Benchmarked model quality with MAE and RMSE.
CO2 Emissions and Fuel Use in Europe
Time-series and regression analysis of European emissions and fuel consumption data.
- Modelled data covering 1990 to 2020.
- Applied regression and ARIMA forecasting.
- Improved out-of-sample accuracy over a linear baseline.
AI Healthcare Validation Analysis
Worked on evaluation workflows for an AI-powered semen analyzer in a healthcare startup environment.
- Compared device outputs against reference measurements.
- Used sensitivity, specificity, and Bland-Altman analysis for agreement assessment.
- Supported FDA regulatory submissions and peer-reviewed publication work.
International Retail Coffee Price Dynamics
Early quantitative modeling project presented to an external examiner panel.
- Modelled international retail coffee price margins as a stochastic random walk in MATLAB.
- Focused on pricing dynamics and quantitative interpretation.
Food System Emissions Explorer
An interactive Shiny dashboard based on my DSS notebook analysis of food-system greenhouse gas emissions. It uses EDGAR, food-product, and GLEAM datasets to make emissions trends explorable in a web interface.