This session presents a novel framework leveraging Large Language Models (LLMs) and Retrieval Augmented Generation (RAG) to create synthetic time-series data for designing automotive braking systems. Mantek’s approach addresses the challenge of limited real-world data by generating realistic simulations of braking performance for new material compositions. Mantek will demonstrate how LLMs can capture complex temporal dependencies and statistical properties of braking system behavior, even when trained on limited data. He’ll compare our approach to traditional time-series methods, showcasing its superior performance in generating high-fidelity synthetic data. Finally, he’ll show how this synthetic data can be used to improve downstream tasks, such as anomaly detection and material selection, leading to accelerated design cycles in automotive manufacturing.