MLBuild enforces inference performance SLAs in CI, automatically blocking slow ML models before they reach production.
