Alternative Credit Scoring ML Model

Overview: Developing a machine learning model to predict creditworthiness for informal sector workers who lack traditional credit histories. We will use alternative data sources like mobile money usage, airtime purchases, and utility bill payments to generate scores. This project aims to expand financial inclusion for thousands of unbanked individuals. Deliverables: - A trained ML model with documented performance metrics (AUC, Precision/Recall). - Data ingestion and feature engineering pipelines for alternative data. - An API endpoint for real-time credit score retrieval. - A technical report on feature importance and model interpretability. Tech Stack: - Python, Scikit-learn, XGBoost, Pandas, SQL. Team Structure: - 1 Lead Data Scientist, 1 Data Engineer.