Optimizing Search and Discovery Systems - Machine Learning Challenges and Solutions

Watch a 41-minute technical presentation from the Data Science Festival Oktoberfest 2024 where Depop's machine learning scientists Odie Edo-Osagie, Ben Ajayi-Obe, and Mitch Watts explore the challenges and solutions in optimizing search and discovery for their marketplace platform. Learn about the team's approach to delivering frequent, high-impact model improvements while maintaining reliability in a cross-functional environment with parallel development. Discover insights into the infrastructure supporting rapid iteration, strategies for managing concurrent development, and methods for effectively deploying experimental features. Gain a high-level overview of how the search and ranking team navigates the complexities of implementing machine learning solutions in a growing marketplace environment.