On the threshold of a new technological era, Sixth Generation (6G) networks promise to revolutionize global connectivity, bringing mobile communications to data speeds in the terabits per second range and ultra-low latency. These networks will enhance the user experience enable a wide range of advanced applications and emerging services. Artificial Intelligence (AI)-powered network functions and services, also known as Network Intelligence Functions (NIF) and Network Intelligence Service (NIS), are essential to achieve this vision. In this study, we present the design and development of an end-to-end framework for orchestrating AI-based functions. Utilizing Kubernetes (K8s) and Prefect, we showcase its implementation through an AI-driven Traffic Classification (TC) use case. Our results confirm the feasibility of the proposed framework, offering valuable insights in the lifecycle management design, such as data collection, decision-making, and critical performance metrics, including deployment time and model performance in terms of accuracy and inference times among three different Machine Learning (ML)-based TC models.

The increasing demand for high-quality and efficient Channel Estimation (CE) in 5G New Radio (5G-NR) systems has prompted the exploration of advanced Deep Learning (DL) techniques. While traditional methods, such as Linear Interpolation (LI) and Least Squares (LS), provide reasonable accuracy and are practical for real-time physical layer processing, recent DL-based CE approaches have primarily focused on accuracy, often without evidence of real-time capabilities. In this paper, we present a comprehensive evaluation of DL-based Super-resolution (SR) methods for CE, comparing models like Super Resolution Convolutional Neural Network (SRCNN), ChannelNet, and Enhanced Deep Super-Resolution (EDSR) in both 1D and 2D convolutional architectures. We optimize these models using NVIDIA TensorRT to reduce computational complexity and latency. Our results show that the optimized 1D-EDSR model achieves the best performance with a Mean Squared Error (MSE) of 0.0126, outperforming all other models in terms of accuracy. However, the optimized 1D-EDSR model fails to meet real-time constraints due to additional computational overhead (0.6798 ms/sample). In contrast, the 1D-SRCNN model offers a balanced trade-off between MSE (0.01738) and inference time (0.0866ms/sample), achieving 40% higher accuracy than LS (0.0288) while maintaining the best energy efficiency (1.48 mJ/sample).