Milan J Parikh’s research and professional interests focus on enterprise data architecture, cloud data platforms, and advanced analytics within the Microsoft ecosystem. His work emphasizes Microsoft Fabric, Azure data engineering, Power Platform governance, Dynamics 365 architecture, and business intelligence using Power BI. He is particularly interested in data governance, scalable data integration, digital transformation, and modernizing enterprise systems to enable data-driven decision making and improve organizational efficiency

Milan J Parikh is a Lead Enterprise Data Architect with over 10 years of experience in enterprise data platforms, analytics, and Microsoft technologies. He currently leads data and analytics architecture at Cytel, where he focuses on modernizing enterprise reporting and data platforms using Microsoft Fabric, Power Platform, Azure, and Power BI. Milan specializes in enterprise data architecture, platform modernization, data governance, and scalable analytics solutions. Throughout his career, has helped organizations optimize costs, improve data accessibility, and implement secure, high-performance data ecosystems. holds a Master’s degree in Electrical and Computer Engineering from San Francisco Bay University and multiple Microsoft certifications, including Microsoft Power Platform Architect and Azure Fundamentals. Reference: Generative AI for ETL Automation: Redesigning Data Flow Management for Heterogeneous Enterprise Environments Modern enterprises manage data across dozens of heterogeneous sources: structured, semi-structured, unstructured, and graph-based. Traditional rule-based ETL pipelines were not built for this reality. Schema drift alone causes significant maintenance overhead and development costs that scale poorly as data complexity grows, particularly when organizations struggle to keep up with evolving data formats and integration requirements. This session presents a generative AI framework that automates ETL pipeline construction and adaptation. The system interprets schema changes dynamically, generates transformation logic without manual intervention, and reduces dependency on rigid template-driven workflows. When compared to traditional ETL methods, the framework shows clear improvements in reducing the time needed to maintain pipelines and better handling of changes in data structure. Attendees will understand where rule-based ETL breaks down, how generative AI addresses those failure points, and what practical deployment looks like in enterprise data environments