With the advancement of mobile communications, Post-5G networks (B5G/6G) will become more complex, resulting in significant challenges for new use case scenarios. Agribusiness, critical mission, and digital twins are strategic examples for Brazil of the use of B5G/6G and are investigated in this project. The complexity generated by the enormous number of parameters and optimizations to be made in the protocol stack to ensure good end-to-end performance brings difficulties to the implementation and management of these networks. The use of artificial intelligence and machine learning (AI/ML) in trend prediction and pattern extraction from the collection of environmental measurements and performance indicators plays a prominent role in B5G/6G networks by allowing configuration without human intervention, with automatic choice of parameters and performance optimization. However, many scientific studies apply AI/ML without guarantees that it will obtain robust results in real scenarios. In fact, the use of AI/ML in mobile communications is still in its infancy, mainly due to the small amount of data available. The methodological strategy of this project overcomes this difficulty through an elaborate collection of data necessary for the design of AI/ML models for the problems studied, adopting simulators and experimental platforms. The main hypothesis is that AI/ML can be applied efficiently in some relevant challenges found in next-generation mobile networks. This project will investigate radio resource scaling, codebook design, physical layer parameterization, and other problems that involve the choice of operating parameters, ensuring the proper functioning of networks and optimizing the use of resources. Based on scientific rigor, this project aims to advance the state-of-the-art by integrating traditional model-based techniques with the use of AI/ML, in order to guarantee adequate solutions to the problems in the prioritized scenarios.