Vaibhav Aggarwal

As data ecosystems evolve, building and operating reliable, maintainable, and scalable data pipelines becomes increasingly complex. This session introduces a modern shift in data engineering: a zero-code ETL platform where users define what the pipeline should do, and data engineers define how the platform should handle its execution at scale. It essentially abstracts pipeline complexities behind an intuitive UI and then standardised configurations. We extend this architecture with an LLM-powered segmentation layer on top of the data warehouse, turning raw data into actionable insights. It converts high-level user intent into SQL queries and downstream pipelines, allowing business users to run experiments on their own—without depending on engineering teams or facing bottlenecks. Zero-code ETL: users see magic, engineers maintain the illusion.