Search papers, labs, and topics across Lattice.
1
3
2
Ditch the reactive autoscaling: a new RL-powered Kubernetes autoscaler learns to anticipate traffic spikes and optimize GPU inference deployments entirely in simulation.