TinyML has made tremendous progress in the last few years, and many tools exist to develop and test machine learning algorithms. Yet, developers face many challenges when it comes to integrating these networks into a complete application. The neural networks become a tiny part of a more extensive system, with the complexity of sharing computing resources and memory, synchronizing events, and integrating with another code base. Deployment is another level of sophistication, with the headache of provisioning the neural network without interrupting the device’s operation. These problems are solved for edge servers and gateways using software containers and orchestration with Docker and Kubernetes. However, these cloud-native solutions are impossible for TinyML applications. Nonetheless, the same concepts are applicable. In this lecture, we will show how using embedded software containers like MicroEJ can simplify the integration and deployment of AI at the tiny edge.