Deep learning computer vision workflow on the Food-101 dataset, from exploratory analysis to deployment, built with an MLOps-first mindset.

• Multi-class image classification on 101 food categories (~101 000 images).
• Systematic benchmark of modern architectures (CNNs and Vision Transformers) under frozen and fine-tuning regimes.
• MLOps stack with MLflow integration, a Streamlit inference application, Docker/Docker Compose and CI/CD with GitHub Actions.
• Reproducible, production-oriented structure (configs, src, tests, pinned dependencies).

Smaller but complete Kaggle competition workflows covering EDA, modeling and interpretation.

• Bank term-deposit classification (binary): Predicts whether a customer subscribes to a term deposit using campaign, demographic and behavioral features, with stratified cross-validation, feature engineering and SHAP-based interpretation.

• Podcast listening time regression: Regression pipeline to predict podcast listening time from episode and show metadata, using K-fold CV, compact model ensembles and diagnostic analysis (residuals, SHAP, calibration).

Interactive R Shiny application organised as an “interactive blog” that illustrates core data science and applied mathematics topics through visual storytelling.

• Covers machine learning themes such as clustering, regression, NLP and time series, plus broader mathematical ideas including Monte Carlo methods, Markov chains and simple epidemiological models.
• Built as a full Shiny app with dedicated data preparation scripts, modular UI/server components and reproducible dependencies managed through renv.