The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Building, Adopting, and Maturing LinkedIn's Machine Learning Platform with Ya Xu - #453
Feb 4, 2021
Ya Xu, the Head of Data Science at LinkedIn, shares her journey from a statistics background to leading a global ML team. She describes the crucial three phases—building, adoption, and maturation—of any successful platform. Ya discusses the challenges of collaboration between application and platform teams, as well as the importance of overcoming 'hero syndrome'. She also dives into LinkedIn's approach to differential privacy, the evolution of experimentation in ML, and how these strategies enhance user experience while ensuring security.
AI Snips
Chapters
Transcript
Episode notes
Shared Accountability
- Make application and platform teams accountable for platform feature success.
- This shared responsibility fosters collaboration and ensures features solve real problems.
Champion Model
- When driving platform adoption, start with enthusiastic early adopters.
- Their feedback helps refine the platform, and their success attracts other teams.
Demand and SMEs
- Encourage application teams to demand platform solutions instead of building their own.
- Establish Subject Matter Experts (SMEs) within application teams to bridge the gap.