Revolutionizing Machine Learning with Netflix's Axion: A Journey in Building a Scalable Fact Store
Thursday, June 6, 2024
Netflix’s journey in enhancing its machine learning capabilities led to the creation of Axion, a revolutionary fact store designed to optimize the quality and accessibility of data for machine learning algorithms. This talk will delve into the evolution of Axion, highlighting its critical role in advancing Netflix’s recommendation systems. We will explore the challenges faced in managing and utilizing large-scale data and how Axion’s innovative architecture addresses these challenges. The session will provide insights into the technical intricacies of Axion, including its four main components: the fact logging client, ETL, query client, and data quality infrastructure. Attendees will learn about the unique approaches employed for data logging, transformation, and querying that ensure efficient and reliable data access for machine learning feature generation.