Build and evaluate personalized text-and-image generation systems for real-world marketing use cases using state-of-the-art GenAI.
Personalization is the engine behind modern digital growth. In this workshop, you’ll learn how leading platforms like Spotify approach preference-aware content delivery—and how generative AI is expanding what personalization can look like across ads, e-commerce, and digital media. We’ll explore how to generate tailored copy and visuals, design retrieval-augmented pipelines, and evaluate impact with user-centric metrics and A/B testing, with a focus on what actually scales in production.
Speaker
Abhinav Kumar
Abhinav Kumar is a Staff Research Scientist at Epsilon, where he has led the development of machine learning models for advertising personalization at scale, impacting billions of users and operating on petabyte-level data. He earned his Ph.D. in Computer Science (2022) from the University of Illinois Chicago, focusing on multimodal conversational agents. His work spans generative AI, multimodal learning, and production-grade personalization systems for LLM-powered applications.
Agenda
Why personalization matters at scale (Spotify case)
How GenAI enables personalized text & image generation
Multimodal personalization: combining text, images, and user signals
Measuring “good personalization”: metrics & A/B testing intuition
From workshop to PBL: what students actually build in the Spotify project
Q&A
Related Project
https://program.blendedlearn.org/pbls/ai-in-marketing-personalization---spotify-project