Deployed, end-to-end ML apps you can open and run — not just slides.
Live Demo
Financial Market Anomaly Detector
Real-time risk monitoring on live market data using an Isolation Forest model, with 7 engineered features (Bollinger Bands, return Z-scores, volume-spike ratios). Interactive Streamlit dashboard with Plotly overlays and a NORMAL / ANOMALY / EXTREME risk table for any ticker.
End-to-end churn prediction on a 7K+ Telco dataset with XGBoost, containerized in Docker and served via a FastAPI REST API. Scikit-learn preprocessing pipeline and a live Gradio interface scoring customers into High / Medium / Low risk.
End-to-end analytics-engineering stack — dbt transformations on Snowflake, orchestrated with Airflow, with Python data pipelines modeling a production-style modern data platform.
Clinical predictive modeling on patient records using a Random Forest classifier to flag at-risk patients across conditions including lung cancer, breast cancer, and opioid overdose.