The presentation by Niranjan Ballal, Thomas Soot, and Dr.-Ing. Michael Dlugosch from Fraunhofer EMI, titled "Generation for Data-Centric AI Applications in Engineering," focused on the vital role of efficient data generation in applying Machine Learning (ML) methods in engineering contexts, particularly for predicting injury risks in multi-modal traffic safety. The motivation behind their research, part of the ATTENTION project funded by the German Federal Ministry for Economic Affairs and Climate Action, is to address the significant challenge posed by vulnerable road users (VRUs) who constitute the largest share of deaths in urban traffic. The research emphasizes the overproportionate share and increasing severity of injuries among VRUs and suggests that ML methods can predict structural dynamics within sub-seconds. However, the complexity and scarcity of training data are major hindrances to the widespread adoption of ML-based injury risk prediction. The presentation stressed the importance of data-centric ML, which focuses on data rather than the ML model to increase efficiency and performance in complex domains. It involves engineering experts to interpret data and its context, ensuring that the data contains all relevant information and formalizing this expertise into rules and metrics. This approach is aimed at generating high-quality synthetic data, leveraging small datasets for ML, and dealing with complexity while increasing explainability. The stages of adaptive data generation were discussed, highlighting the use of existing simulation data from former engineering R&D processes, deliberate data generation with Design of Experiments (DoE) methods, and using feedback loops to optimize ML model performance by iteratively resampling the most valuable data points. The goal is an adaptive data generation pipeline that efficiently optimizes for performance and reusability. The pipeline's implementation and testing for both generic mathematical and real application scenarios were mentioned, along with its use in the ATTENTION project. Future work includes evaluating the overall pipeline performance, introducing further expert-driven metrics, assessing the effects of different ML model types, and adapting and evaluating more use cases.