🫀 Excited to share my latest ML project — Heart Disease Predictor!

I built an end-to-end machine learning application that predicts the presence or absence of heart disease based on patient vitals like blood pressure, cholesterol, chest pain type, and more. The dataset is taken from a kaggle competition ( playground Season 6 episode 2) which i participated in. 🔍 Problem it solves: Early detection of heart disease risk is critical. This tool allows anyone to input patient data and instantly get a prediction along with a visual explanation of which factors contributed most to the result — making the model transparent and interpretable, not just a black box.

⚙️ Technologies Used: • Python • XGBoost — for high-accuracy classification • Optuna — for hyperparameter tuning (achieved ~95.5% AUC) • SHAP — for explainable AI / feature importance visualization • Scikit-learn — for preprocessing and evaluation • Joblib — for model serialization • Streamlit — for building and deploying the web app • GitHub — for version control and deployment

📌 Key highlights: ✅ Tuned XGBoost model with Optuna over 25 trials ✅ SHAP values to explain every single prediction ✅ Clean, minimal UI deployable in one click ✅ Fully deployed and accessible via a public URL

This project taught me how to take a model from a Jupyter notebook all the way to a live, interactive web application — a skill I believe every ML engineer needs.