Trimming the Long Tail of Production Model Ownership at Hinge

In this 29-minute talk from Data Council's 2025 Day 2 MLOps & Platforms session, explore how Hinge approaches model serving as a long-term ownership challenge. Staff Machine Learning Engineer Jonathan Jin shares valuable insights on focusing on the "long tail" of model ownership to create sustainable AI platforms. Learn strategies for building effective model serving infrastructure and addressing observability challenges while creating a "golden path" that enables AI teams to innovate and scale modeling investments without operational burden. Gain practical perspectives applicable whether you're developing MLOps infrastructure or working to improve the sustainability of AI systems in production.