Due to the limited hardware resources of airborne edge devices, rapid deployment of AI models on them while taking into account the versatility of AI model porting and the diversity of tasks is an urgent problem to be solved. This paper presents a generalized light intelligence framework for airborne MCU-edge devices, which uniquely combines an AI inference component with an IoT OS to manage and port AI applications while taking care of multi-task management, energy consumption. Firstly, the AI network model is converted into AI C files that can be run on embedded devices by an AI inference component such as X-CUBE-AI or Huawei msmicro tool. Then, the lightweight IoT operating system (IoT OS) such as FreeRTOS or Huawei LiteOS is deployed on the MCU to provide necessary services such as connectivity, security and task management. In addition, the converted AI C files are ported as a stand-alone part of the IoT OS to form an IoT OS integration project, which is finally deployed on the MCU. Experimental results show that compared to deploying the AI model with a separate AI inference component, the inference time is almost unchanged after using the proposed light intelligence framework, although the storage footprint is a bit higher. Moreover, the energy consumption increases by about 15%, which is acceptable loss considering the system management benefit of the IoT OS. To summarize, the light intelligence framework enables rapid deployment of AI applications and collaborative multi-task management in low-power mode on airborne edge devices.