Umar Faruk

I built and deployed end-to-end machine learning applications for telecom churn prediction, Fashion MNIST image classification, and IMDB sentiment analysis using Python, Scikit-learn, PyTorch, Streamlit, and WEKA.

My churn work analyzed 7,043 customers, achieved up to 86.14% ROC AUC, identified contract type and tenure as key churn drivers, and produced capacity-aware retention recommendations. I also analyzed 9,994 retail transactions across Excel, Power BI, and PostgreSQL, building dashboards and validating KPIs across all three tools.

Alongside my Chemical & Petroleum Engineering degree, I gained industrial experience at Saro Agro Sciences Lifecare, Dangote Cement Plant, and the National Inland Waterways Authority, with exposure to production, quality assurance, R&D, maintenance, and safety operations.