Does the Kubernetes API Server use leader election like the Scheduler does?

No — every API Server replica is independently active and serves requests simultaneously, since handling read/write API requests doesn't have the 'only one instance should be doing this at a time' requirement that scheduling and reconciliation do.