In recent years, the rapid advancements in Artificial Intelligence (AI) technologies and the increasing pervasiveness of Internet of Things (IoT) applications have made embedded platforms more complex to design and implement. Such systems are tightly coupled with the environment, that is continuously sensed to run the algorithms on fresh data, communicate the generated results to any external actors, and react accordingly through actuating features. This implies that, depending on the application, constraints that go beyond pure performance, such as response time, power consumption or safety, may be imposed, therefore requiring a greater amount of development time and resources to build the complete system.