Modern automotive systems generate massive, heterogeneous data streams spanning engineering design, manufacturing operations, connected vehicle telemetry, warranty claims, service records, and customer feedback. Converting this multimodal data into actionable intelligence remains a fundamental challenge because of scale, fragmentation, temporal dynamics, and the coexistence of structured and unstructured information. In practice, enterprise AI systems must reason over billions of sensor events, hundreds of millions of repair records, and petabyte-scale engineering repositories distributed across global manufacturing and service ecosystems. This challenge is compounded by high-velocity streaming data, evolving vehicle configurations, noisy operational signals, and the need for low-latency inference, traceability, and enterprise-grade reliability.

This talk presents a unified portfolio of industrial AI systems deployed across the global vehicle development lifecycle, showing how machine learning, deep learning, and generative AI can be integrated to deliver large-scale engineering and operational intelligence in complex cyber-physical environments. Key production-scale applications include: (1) automation of Failure Mode and Effects Analysis (FMEA) for proactive failure mode identification; (2) device-level analytics for manufacturing assets using high-frequency multisensory time-series data for anomaly detection; (3) early fault detection using diagnostic trouble codes and telemetry signals; (4) emerging issue detection through multimodal fusion of structured and unstructured enterprise data; (5) issue retrieval, insight extraction, and read-across across engineering, manufacturing, quality, and service repositories; and (6) automatic repeat-repair detection and recall-effectiveness analysis. Collectively, these deployed AI systems have delivered more than $400 million in business impact year over year.

Beyond these application case studies, the talk highlights core research challenges in deploying foundation-model-driven AI in industrial settings, including representation learning over heterogeneous data, scalable retrieval and reasoning, robustness, governance, and continuous adaptation under evolving operational conditions. Together, these systems illustrate how modern AI architectures can transform automotive engineering ecosystems by enabling earlier fault discovery, faster root-cause analysis, and more data-driven decision-making across the vehicle lifecycle.