The Cognition team needed better autoscaling, availability and features for our GPU workloads. The Common Platform Team was there to help.