Gurleen Kaur Kailey

Marketing Campaign Optimization:

• Collaborated with marketing teams to craft data-driven strategies, segmentation models, and targeted campaigns.
• Analyzed KPIs like customer acquisition costs, conversion rates, and ad ROI to enhance marketing efficiency, optimize spend, and improve campaign effectiveness.

Loan Volume Forecasting:

• Developed and managed a predictive model for month-over-month loan volume forecasting, providing critical insights to the liquidity management team. Focused on accurately estimating loan deltas to support informed decision-making in liquidity planning.

Customer 360 & Forecasting Improvements

• Led a Pharmacy 360 solution to deepen customer insights and personalize offers, enhancing segmentation and marketing effectiveness.
• Improved demand forecasting accuracy by replacing the fbprophet model with a multi-output Random Forest regressor, reducing MAPE by 15% and speeding up processing through parallel execution.
• Developed an automated offer generation solution in Databricks, optimizing workflows with reusable notebooks and streamlined data processing.

Assortment Dashboard & Recommendations

• For a retail client, led the end-to-end creation of an Assortment Enablement dashboard in PowerBI, collaborating with strategy and category teams to drive a 15% increase in sales through data-driven insights.
• Implemented a recommendation algorithm to optimize product mix, achieving a 5% uplift in average units per category.
• Developed an outlier detection model in Python to improve A/B testing , identifying non-parallel trends in 20% of cases to refine decision-making accuracy.

PII Detector: Technology Client (US)

• Worked in the role of an ML Developer for a major cloud service provider to enhance GDPR compliance. The project evolved from a regex and Power Automate solution to an efficient ML-driven approach for PII data masking.
• Utilized Kusto Query Language (KQL) to seamlessly transfer data from Azure Data Explorer to Azure Storage, ensuring security and accessibility.
• Developed and deployed SVM model for PII classification. Achieved significant improvements in classification accuracy with a precision score of 85% and recall score of 90%.