Rohit Sivaprasad, Head of Data Science and Engineering at Adobe Stock, joins Roy Daniel to discuss what it takes to prepare legacy data platforms for the AI era. The conversation explores how organizations can untangle years of ad hoc business logic in data warehouses, reduce accumulated technical debt, and build the semantic foundations required for AI agents to reliably access and use enterprise data.
Drawing on experience from Adobe, Walmart, AWS, and Meta, Rohit explains the early architectural patterns emerging in agentic data engineering and the practical steps teams are taking to modernize data platforms. The episode also looks at the specific challenges of supporting Adobe Stock’s hybrid B2B and B2C business model, where subscription and ecommerce workloads place different demands on the data platform.