the work

Projects

A few things I’ve built, explored, and brought to life.

Customer Intelligence Engine

PythonScikit-learnKMeansFastAPIDockerAWS EC2

Built a KMeans-based customer segmentation system that classifies customers into High-Value and Budget personas using seven behavioural features.
Validated segmentation using the Elbow Method and Silhouette Score to determine meaningful customer clusters.
Containerized the application with Docker Compose and deployed it on AWS EC2, exposing the model through a real-time API.

PythonCatBoostFastAPIDockerAWS

Built a production-grade bike demand forecasting pipeline using CatBoost to model demand patterns.
Used time-based train/test splitting and hyperparameter tuning to validate forecasting performance without temporal leakage.
Containerized and deployed the complete system on AWS, serving real-time predictions through a live dashboard.

PythonPandasMySQLPlotlyStreamlitRailway

Built a SQL-based ABC/Pareto classification engine identifying 93 Class A SKUs responsible for 79.7% of revenue.
Used the analysis to turn raw inventory data into actionable product and revenue classifications.
Debugged duplicate ingestion and JOIN amplification issues, restoring data accuracy across 9,001 records.

more projects

A few more things I’ve explored along the way.

End-to-end product analytics project for a B2C digital wallet, exploring user behaviour, engagement, and business decisions through data.

A small interactive Python desktop experiment that reacts to keyboard activity and brings a playful virtual cat to life.

An end-to-end Power BI analysis exploring taxi data through interactive dashboards and business-focused visualisation.