Software and functionality previously contained in the cloud is becoming firmly entrenched in the network edge. However, most orchestration software is still heavily focused on cloud architectures and centralized control, whereas the edge is a highly decentralized concept. Moreover, useful orchestration parameters are different in the edge than they are in the cloud, while software security must be tackled from additional perspectives. This project proposes several research topics to enable a secure, decentralized “bottom-up” edge orchestration approach, firstly by enabling edge devices to determine optimal service providers by discovering their neighborhood, and matching discovered services to user requirements using an intent-based framework. Furthermore, device discovery and service orchestration will be assisted by novel AI algorithms and optimized neural network models, while new approaches for decentralized learning will extend flexible computation and orchestration all the way to microcontrollers. Finally, the project also aims to use device attestation and a newly developed approach for decentralized container networking to create a secure environment for edge service execution and communication.