Mmopiemang
Mmopiemang

A Python project that builds telemetry points for both drivers and circuits, then predicts how a driver would likely perform on a track they've never raced. This comes with a plain-language explainability layer showing why.

A framework-agnostic Python package, CLI, and FastAPI service that audits ML models for 5 common failure modes: class imbalance, small sample instability, target leakage, temporal leakage, and data drift. It returns a plain-language report card with structured evidence instead of a bare pass/fail.

Used macro and micro data from GSMA and the World Bank Index to predict formal saving rates for informal traders. Built an ensemble modelling pipeline combining Random Forest, Gradient Boosting, and Regression, with visualisations in Python via Bokeh and Panel.

Designed and built a full-stack voting portal for a local hip-hop show as a freelance MVP, implementing RBAC and Python-based authentication for secure vote management.

Academic research paper evaluating the impact of Business Intelligence tools in measuring ESG performance and accountability within the automotive sector.