Python ML Developer · Financial Risk & NLP Systems · Streamlit Apps
B.Tech Computer Science (Data Science) · Lovely Professional University · 2029
I build end-to-end machine learning systems focused on financial risk, NLP, and explainable AI — fully deployed as live Streamlit applications.
- 🎯 Specializing in fraud detection, credit risk modeling, and sentiment-driven stock analysis
- 🔍 Strong focus on model explainability — every model I ship includes SHAP analysis
- 🚀 All projects are production-deployed, not just notebooks
- 💼 Open to freelance ML/data science projects on Upwork
| Project | What It Does | Key Tech | Live App |
|---|---|---|---|
| 🔍 Fraud Detection + AI Explanation | Detects credit card fraud (0.978 ROC-AUC) and explains why in plain English | XGBoost · SHAP · GPT-3.5 · SMOTE | ▶ Demo |
| 📈 FinSentiment Stock Predictor | Live news sentiment + technical indicators → next-day price movement prediction | FinBERT · XGBoost · NewsAPI · yfinance | ▶ Demo |
| 💰 Loan Default Risk Engine | Predicts borrower default probability on 1.3M+ records with 0–100 risk scoring | XGBoost · SHAP · Lending Club data | ▶ Demo |
| 🏢 AI Bankruptcy Risk Analyzer | Predicts corporate bankruptcy up to 5 years in advance (0.97 ROC-AUC) | XGBoost · SHAP · Polish Companies dataset | ▶ Demo |
| 🛒 Walmart Sales Forecasting | 4-week department-level sales forecasting using recursive multi-step XGBoost | XGBoost · Time-series FE · Streamlit | ▶ Demo |
- 🔧 Expanding explainability features across all deployed projects
- 📊 Adding SHAP dashboards and training notebooks to each repo
- 💼 Actively taking on freelance ML projects via Upwork
All projects are live and deployed — click any demo link above to try them.