🔭 Currently: Building end-to-end data solutions and expanding my MLOps expertise
🌱 Learning: Advanced ML techniques, AWS deployment, and Power BI analytics
💬 Ask me about: Predictive modeling, data storytelling, and Python automation
⚡ Fun fact: I approach data problems like detective cases - following clues to uncover insights
🛠 Technologies: XGBoost · SMOTE · Feature Engineering · Model Evaluation
- 📊 Applied SMOTE to balance classes and boost recall for minority churn cases
- 🔍 Identified key churn drivers using feature importance analysis
- 🚀 Achieved ROC-AUC of 0.89 with robust cross-validation
🛠 Technologies: Python, Random Forest, SHAP, Gradio
- 📈 Achieved 97.5% R² score in predicting power demand
- 🔎 Built interactive dashboard with anomaly detection
- ⚡ Enabled strategic energy planning for cost reduction
🛠 Technologies: Gradient Boosting, Power BI, Feature Engineering
- 🏛️ Identified key cost drivers (Municipal Budgets 47.8%, GDP 22.1%)
- 📊 Achieved 0.681 R² score with robust cross-validation
- 📈 Created executive dashboards for budget planning
🛠 Technologies: Python, LightGBM, Gradio, Power BI, Feature Engineering
- 🔒 Achieved 97.04% accuracy in real-time fraud detection
- 📊 Developed interactive dashboards & KPIs for executive insights
- 🌍 Enabled geographic, device, and customer risk analysis for prevention strategies
- 🚀 Deployed live on Hugging Face → FraudGuard AI Demo
