By leveraging the modern machine learning algorithms, we can build up more Artificial Intelligence (AI) systems, like self-driving cars, smart factories and financial analysis systems, to improve our daily life. In addition to building up an AI system, several prerequisites are required to drive the system, including data collection, data storage, machine learning models, training dataset, parameters tuning, and so on. To obtain the benefit of scalability and flexibility, most AI systems are built on a cloud platform, which shares resources with others in the same infrastructure. Though the above concept is trivial, the implementation faces big challenges when realizing it. In this paper, an easy-to-use cloud framework for machine learning as well as its implementation guideline is presented for building up a cloud-based development platform. We conduct several experiments on analyzing and monitoring the health condition of bearings of motors. We compare and analyze the feasibility of the proposed framework.