azhimako1702

Developed and implemented machine learning models to detect fraudulent transactions. Regularly reviewed and refined models to ensure ongoing reliability and accuracy in fraud detection. Built a Levenshtein-based name similarity scoring module to prevent high-risk transactions linked to terrorism watchlists or politically exposed persons (PEPs). Adapted to the dynamic needs of the organization by undertaking different ad-hoc tasks, demonstrating flexibility and problem-solving skills.

Developed segmentation and uplift models to optimize collection strategy and improve recovery performance. Engineered SQL-based feature pipelines for predictors affecting payment probability. Automated dashboards in Power BI / Looker for field & call treatment performance tracking. Supported scorecard modeling and control-group analysis for credit-risk experimentation. Supported ad hoc analysis related to multiple processes in collections.

Designed and validated digital-banking workflows embedding risk and compliance logic. Collaborated with engineering and QA to ensure data integrity via SQL-driven validation. Defined functional specs aligning system behavior with regulatory requirements.

Defined fintech business requirements and risk scenarios for new platform modules. Utilized SQL for backend logic testing, identifying potential system gaps during UAT. Collaborated with compliance teams to integrate regulatory parameters into system design.