Computer Vision & ML
Loan Eligibility Prediction System
Engineered an end-to-end ML pipeline with feature engineering, SMOTE oversampling, and GridSearchCV for loan approval prediction, deployed as a RESTful API serving 1000+ daily requests.
Key Architectural & System Features
Engineered ML pipeline using CIBIL score, income, loan amount, and credit-history features
Trained and compared Logistic Regression, Random Forest, and XGBoost classifiers
Achieved 99.5% accuracy and deployed as FastAPI service for production workloads
Technologies & Tools Stack
Scikit-learnXGBoostFastAPIDockerPostgreSQL