How Torc hit 90% GPU utilization and other stories on scaling AI with Ray from Discord, Cubist, and Coinbase

This article recaps presentations from Ray Day NYC featuring case studies from Discord, Cubist, and Coinbase on scaling AI workloads using the Ray distributed computing framework, with a focus on Torc's achievement of 90% GPU utilization optimization. The content covers practical techniques and lessons learned for efficiently distributing AI/ML training and inference across large clusters using Ray's scaling capabilities.

Source: Anyscale/Ray | 2026-06-10