• Trained, deployed and monitored 10 machine learning models for predicting customer behaviour including churn and mobile upgrade. Consistently ¬achieving AUC of above 0.7 on highly imbalanced datasets.• Trained NLP models classifying topic and sentiment for many data sources including X posts, Facebook comments, customer survey responses, employee HR queries.• Lead the end to end development of a data pipeline for training and deploying over 40 time series models to predict network KPIs. Predictions presented on Cognos and Power BI dashboards for consumption.• Built and deployed 3 external customer facing dashboards in Tableau for use in the data monetization strategy. Performed extensive data checks on the underlying data and presented the dashboards to product owners within the business.• Technologies: Python, Teradata SQL, Pandas, NumPy, Scikit-learn, Statsmodels, Matplotlib, Jupyter, Git

• Listed as a co-inventor on a patent titled “AUTOMATIC HOT AREA DETECTION IN HEAT MAP VISUALIZATIONS”. I proposed the chosen clustering algorithm (DBSCAN) and tested our solution on matrices containing arbitrary shapes of hot areas. U.S. Patent 11727030 granted August 15, 2023.• Designed automated web browser tests with Python Selenium to test the User Interface of SAP Analytics Cloud.• Worked under the scrum framework of agile software development.