Just completed my GPU load balancer that's smart enough to estimate how long your AI workloads will take balances and optimizes which queues to use.

Very gm indeed