I build ML systems that turn "we lost a sale" into "we saw it coming 4 days ago."

I help e-commerce and retail businesses stop losing money to stockouts, dead stock, and customer churn — using forecasting and ML systems that tell you exactly what to do, not just a number on a dashboard.

🥈 2nd Place — Data Analysis, Technofest Competition (Jan 2026)

Deployed demand forecasting system that catches stockouts before they happen and frees up cash trapped in slow-moving stock. XGBoost model cuts forecast error by 39.5% vs. baseline, flags critical stockouts 4 days in advance, and identified $4,275 in capital tied up in overstock — fully automated with nightly retraining via GitHub Actions. Stack: Python, XGBoost, Streamlit, GitHub Actions · Live demo

Predicts which customers are about to leave, explains why using SHAP, and recommends the exact retention action to take. Targeted retention strategy shows an estimated 60% cost reduction ($210K saved) vs. blanket offers, with 75% recall on at-risk customers. Stack: scikit-learn, XGBoost, SHAP, MLflow, Optuna, Streamlit · Live demo

Power BI revenue intelligence dashboard for a hotel group, turning raw booking data into actionable insights on occupancy, revenue, and performance trends. Stack: Power BI, DAX, Data Modeling

End-to-end data analysis pipeline — from raw SQL scripts to Python transformation to a Power BI dashboard — analyzing e-commerce performance across 100K+ orders. Stack: SQL, Python, Power BI

Languages & Core: Python, SQL Data Analysis: Pandas, NumPy, Matplotlib, Seaborn, EDA ML / Deep Learning: scikit-learn, XGBoost, CatBoost, Deep Learning, Feature Engineering MLOps: MLflow, Docker, GitHub Actions, DVC Dashboards & BI: Power BI, Streamlit Backend: FastAPI, Streamlit Data Extraction: APIs, Databases

I work with e-commerce and retail businesses to build forecasting, churn prediction, and analytics systems that pay for themselves. Open to project-based work.

📩 Reach me: haseeb16001@gmail.com