This conference talk presented at All Things Open AI 2025 by Dhivya Nagasubramanian from U.S. Bank explores the methodical process of scaling Generative AI projects from proof-of-concept to enterprise-level implementation. Learn how to adopt an incremental scaling approach that builds on lessons from each development stage to create solutions that meet real-world requirements. Discover techniques for refining models based on early user feedback, optimizing performance throughout the development lifecycle, and strategically expanding capabilities. The presentation covers practical strategies for testing in controlled environments, leveraging real-time data collection, and implementing agile methodologies to minimize risks during scaling phases. Walk away with a framework for creating robust, adaptable GenAI solutions that can deliver significant impact at scale through thoughtful iteration rather than rushed deployment.
Overview
Syllabus
Scaling Success: Transforming PoC into Large-Scale Impact through Iteration - Dhivya Nagasubramanian
Taught by
All Things Open