Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn how Toyota's Woven division developed AutoTriage, an AI video agent that revolutionizes automated driving system testing by automatically identifying software errors and their root causes from thousands of hours of real-world driving footage. Discover the mission behind achieving zero traffic accidents through advanced automotive software development and understand why automated driving technology is crucial for road safety. Explore Toyota's comprehensive approach to accident prevention and gain insights into the three core components of autonomous driving systems: perception, planning, and control. Examine the key bottlenecks in autonomous vehicle development and see how generative AI addresses the costly and time-intensive manual review process of driving test videos. Navigate the technical challenges of building a video AI agent, including data quality requirements, prompt engineering, and model selection. Understand how the team overcame initial failures through iterative improvements in datasets, prompt design, and model architecture, ultimately evolving from a single-agent system to a sophisticated multi-agent judging framework. Master the four key principles for successful AI agent deployment: creating high-quality datasets, fostering close collaboration with domain experts, maintaining persistence through failures, and conducting systematic experiments. Gain practical knowledge for transitioning AI agents from prototype to production while contributing to the ultimate vision of safer roads through AI-driven automation in automotive software development.
