Safe City-Machine Learning takes the concept of safe city one step further by introducing artificial intelligence to it. Human eye is susceptible to error and human beings while monitoring CCTV cameras can be prone to miss a problematic event. Thus we propose a system where the images from the CCTV camera are taken and a prediction model evolved on previous such images is used to detect further mishaps. The model is continuously evolving to further improve its detection capability. Safe City-ML will incorporate smart machine learning management to detect an event automatically from many images. The system working for a safe city will automatically generate alerts for an operator with the image of the problematic area who then calls on the concerned authorities and nearby law enforcement vehicles in case of any law and order or other situation requiring attention An intelligent transportation system to curb congestion and make inner city roads safe and protected will also be incorporated through sensor monitoring along with CCTVs for scheduling the traffic. Safe City-ML will also have safe homes, safe schools, safe colleges, safe government institutions where continuous monitoring will be done through sensory network, automated reports on various malicious activities and events will be directed to the central and regional monitoring stations.