In an ADAS, high-performance CPUs are the central core for decision-making and other high-level system management. To enable various self-driving capabilities, highly powerful accelerators (DLA, GPUs etc) are used. In this talk, we will first focus on the various AI/ML-based ADAS tasks and the underlying hardware used to execute those tasks. The talk will also focus on the fault tolerance and robustness aspects of these systems. Finally, a brief discussion on how commercial software like Altair EMBED and KnowledgeStudio can be used in the domain.