The goal of the Twitter Home Timeline is to surface the most relevant content to our users. While we have access to lots of implicit data about how users interact with the content on the platform, it is challenging to define an objective function for our models that can be measured in the shot-term. In this talk, I will describe how we use a combination of large-scale machine learning, hypothesis-driven product development and A/B testing to improve the product iteratively while also enhancing our understanding of the objective function.
PARAG AGRAWAL
Parag Agrawal is a Principal Software Engineer at Twitter, focused on improving the user experience in the Home Timeline through relevance. He also works on making better use of A/B testing and data analysis to guide product development for the company. Parag received a PhD in Computer Science from Stanford where he worked on problems related to large scale data management systems.
