Personal website and portfolio: pharmacometrics, PK/PD modelling and data science.

Four self-contained ML projects: fraud detection, demand forecasting, semantic similarity, and an implicit-feedback recommender. CPU-only, tested, built around the evaluation mistakes that flatter offline metrics.

Unsupervised clustering of gene expression and ML prediction of SSRI remission in depression (R, tidymodels) with an interactive results dashboard

Reproducible, self-contained analysis dashboard built from synthetic metabolomics data. No real data anywhere.

Interactive Shiny app for simulating IV dosing regimens from a generic two-compartment population PK model. Drug parameters, covariates and variability are all inputs.