KServe (formerly KFServing) is a Kubernetes-based inference solution for deploying and managing ML models at scale. UnitedLayer integrates KServe in G3 AI Cloud to provide scalable, serverless inference with built-in autoscaling and logging.