A microgrid can be characterized by its integration of distributed energy resources and controllable loads. Such integration brings unique challenges to the microgrid management and control which can be significantly different from conventional power systems. Therefore, a conventional energy management system (EMS) needs to be re-designed with consideration of the unique characteristics of microgrids. To this end, we propose a microgrid EMS named a microgrid platform (MP). We take into account all the functional requirements of a microgrid EMS (i.e., forecast, optimization, data analysis, and human-machine interface) and address the engineering challenges (i.e., flexibility, extensibility, and interoperability) in the design and development of the MP. Moreover, we deploy the prototype system and conduct experiments to evaluate the microgrid management and control in real-world settings at the UCLA Smart Grid Energy Research Center. Our experimental results demonstrate that the MP is able to manage various devices in the testbed, interact with the external systems, and perform optimal energy scheduling and demand response.