TU Darmstadt (TUD) contributes to the ACES project by analyzing security risks and developing novel approaches to ensure the security and robustness of ACES systems against the ever-increasing cyber threats in edge-cloud services. Our innovative solutions have been published and presented at top-tier security conferences, including IEEE S&P, USENIX Security, and NDSS, as listed in the publication section. Notably, two of our papers received Distinguished Paper Awards at the NDSS Symposium 2025—one of the top four security conferences worldwide. In particular, TUD has developed KubeFocus, a comprehensive anomaly detection solution specifically designed for containerized environments such as Kubernetes. Given the high risks associated with container and Kubernetes deployments—including threats like malicious or insecure container images, Docker API abuse, third-party application exploitation, and shadow worm attacks—our solution integrates advanced dynamic deep-learning-based anomaly detection techniques. This approach enhances the resilience of Kubernetes and container systems by enabling real-time threat detection and response. Furthermore, TUD is also developing security mechanisms for distributed machine learning systems, such as federated learning and split learning, to defend against data and model poisoning attacks, as well as inference attacks. This work focuses on privacy-preserving techniques and improving resilience against adversarial threats, ensuring the integrity and security of ML applications in edge-cloud environments. As part of the ACES project, TUD is collaborating with partners to integrate KubeFocus, our advanced anomaly detection system for Kubernetes into the ACES system, strengthening its security framework.