Creating Deeper and Safer Connections with GenAI at Tinder

How do you keep 47 million people swiping, matching, and chatting while guarding against bad actors and making every connection feel personal? In this DeployCon session, Tinder’s ML engineering team peels back the curtain on the dating app’s approach to Generative AI with a deep dive on two critical use cases for safety and personalization. • Trust & Safety at Scale: how the team built an LLM-powered workflow that correctly identifies the signals of bad actors across 47 languages. You'll hear how they went from 33% precision with a rules-based system to nearly 100% precision by fine-tuning sm…