Predictability is crucial in real-time computing systems, as it is fundamental for ensuring consistent, reliable execution within specified time constraints. Without predictability, even occasional delays or unforeseen behaviors can compromise system performance and safety, which may be unaffordable or catastrophic in many time-critical applications. However, artificial intelligence (AI) tasks and high-performance computing (HPC) hardware architectures often introduce unpredictability in program execution due to their inherent complexity and nondeterministic dynamics. Consequently, managing and mitigating unpredictability becomes a key challenge when integrating and leveraging AI or HPC advancements in real-time systems. This project will address this challenge by developing methods and techniques to preserve time predictability in systems, even when traditionally predictable aspects become unpredictable. This project will produce new system models, scheduling algorithms, analysis frameworks, prototypes, and tools. These outcomes will establish a solid foundation for designing and implementing real-time systems that are highly functional, resource-efficient, and predictably reliable. This project will serve as a cornerstone for the design, implementation, and certification of next-generation real-time systems, overcoming the limitations of traditional predictability assumptions. These advancements will be pivotal for sectors such as autonomous vehicles, industrial control