Return of the Redi: Redis and Ray in the ML Ecosystem

In the past year, Ray has made so many additions including changes they made to the store cluster metadata, which was originally based on Redis. As the use cases have expanded to ML based worklaods, Ray and Redis can be used together in ML pipelines, and this reviving the relevance of Redis to the Ray architecture. From simply backing up cluster metadata to pulling real-time features from an online feature store, Ray can utilize Redis in several ML contexts. This talk will discuss the return of the Redi(s): how Redis and Ray can be used together in different types of ML pipelines.