MLOps for Ad Platforms - Lessons from Ad Tech - Coffee Session 130

Explore the intricacies of MLOps in ad platforms with Andrew Yates in this 49-minute podcast episode from MLOps Coffee Sessions. Gain insights into designing ML components within larger systems, the importance of stable interfaces, and the need for intermediate ground-truth signals in ad tech. Learn about the critical balance between profitability and accuracy in advertising, and how it impacts a company's longevity. Discover Andrew's extensive experience leading ads ranking, auction, and marketplace teams at major tech companies, and his expertise in designing billion-dollar content marketplaces. Delve into topics such as the evolution of adtech, team structures in big tech companies, strategies for managing technical debt, and the engineering challenges faced by smaller teams. Understand the complexities of real-time streaming in ad platforms and the potential for companies to integrate their models into systems like Promoted.ai. This comprehensive discussion covers everything from the basics of adtech to advanced MLOps strategies, making it valuable for both newcomers and experienced professionals in the field.

[] Introduction to Andrew Yates and takeaways
[] Want more like this episode?
[] Andrew's Background
[] How did he get into adtech?
[] Evolution of adtech
[] Challenges they face
[] The structures of teams in bigger tech companies
[] Search and discovery teams in bigger tech companies
[] Strategy around technical debt
[] Promoted.ai for big marketplaces
[] How Andrew fits into teams
[] Engineering challenges when working in a small team
[] How much white-gloving they do amid complexity
[] Allowing companies to plug in their models into Promoted
[] Drawbacks with doing real-time streaming
[] Wrap up