_1722956449156001fz1R.png "Anyscale Logo")
eBay's AI platform has undergone a significant transformation aimed at accelerating AI application development, and optimizing expensive GPU resources, and streamlining the user experience. The integration of Ray into our ecosystem has been pivotal in addressing these challenges, leading to a more agile and efficient AI workflow.
In this talk, we will outline the journey of integrating Ray into eBay's AI platform. We will discuss the application of Ray in four key production scenarios: batch inference, near-real-time (NRT) feature engineering pipelines, distributed training, and comprehensive AI solutions. For each of these production scenarios, we will talk about the specific problems we faced, the practical solutions we found and the lessons we learned along the way.
Concluding the talk, we will explore the product design improvements driven by Ray. A well-crafted product design is crucial for providing users with a seamless experience in managing their assets and utilizing the platform throughout the entire machine learning lifecycle.
In this talk, we will outline the journey of integrating Ray into eBay's AI platform. We will discuss the application of Ray in four key production scenarios: batch inference, near-real-time (NRT) feature engineering pipelines, distributed training, and comprehensive AI solutions. For each of these production scenarios, we will talk about the specific problems we faced, the practical solutions we found and the lessons we learned along the way.
Concluding the talk, we will explore the product design improvements driven by Ray. A well-crafted product design is crucial for providing users with a seamless experience in managing their assets and utilizing the platform throughout the entire machine learning lifecycle.
