I developed an end-to-end telecom customer churn prediction pipeline, covering data cleaning, exploratory analysis, feature engineering, and model development. Using XGBoost and SMOTE, I achieved 95% accuracy, then deployed the model as a Flask API with a simple interface for real-time predictions.
At Recykal - EPR Sustainability Solutions, I assisted the Compliance Team with sustainability documentation, regulatory workflows, data verification, and audit preparation. I also taught computer applications and digital tools, and built a banking risk analysis project with classification models and interactive Power BI dashboards.
