Infrastructure Evolution
Chris discusses the evolution of their infrastructure from a homegrown system to Kubeflow, highlighting the challenges of finite memory resources and the need for flexibility in model deployment. The transition to Kubeflow on a new Kubernetes cluster has brought numerous advantages, including improved scalability and support for diverse libraries like fasttext.In this clip
From this podcast

Gradient Dissent - A Machine Learning Podcast
Chris Albon — ML Models and Infrastructure at Wikimedia
Related Questions