Parthiban M_

Engineered a customer-facing GLM pipeline using SAS and Python to automate frequency-severity modeling, risk factor analysis, and premium rate derivation, with interactive dashboards via SAS Visual Analytics for real-time pricing decisioning. Delivered a 70% reduction in manual computational effort, enabling faster and more consistent premium pricing at the point of underwriting for end customers.

Bank customer data was riddled with duplicate and inconsistent records within Snowflake, posing significant data quality and regulatory compliance risks across the enterprise. Built a customer-facing self-service interface using HTML, CSS, and JavaScript enabling bank customers to directly update their information, integrated with SAS Intelligent Decisioning to execute in-database deduplication within Snowflake through fuzzy logic and business rules-driven matching for data governance. Established a governed single source of truth across the customer database, eliminating redundant records at the database layer and ensuring high-integrity, audit-ready customer data in compliance with financial data standards.

Managing and versioning large-scale customer data across distributed sources lacked a unified cataloging and lineage tracking mechanism within the SAS ecosystem. Integrated DuckLake with SAS Viya to automatically retrieve and catalog table metadata, persisting the catalog information into AWS S3 to enable structured, version-controlled data storage within a unified SAS platform. Delivered end-to-end data versioning and catalog visibility for customer data assets, improving traceability, governance, and accessibility across the enterprise data platform.