Level four autonomous technology (automated self-driving with no human driver needed) holds great promise not only for revolutionizing mobility, but also for significantly reducing traffic fatalities. Annual traffic fatalities in the U.S. consistently number higher than 30,000 and are on the rise. Studies state that greater than 90% of crashes are due to driver error, which leads to the optimism for improved safety with autonomous technology. But level four autonomous technology has not yet arrived and may be years away. The engineering challenge of putting this technology into production is considerable. The difficulty can be related to the average fatal crash rate for U.S. drivers which is approximately one per 100 million miles of driving. Verifying a self driving vehicle to a factor better than this number is a key hurdle to realizing the safety benefits. This presentation will describe an approach to develop a safety-specific level four system with significantly reduced verification hurdles, allowing shorter time to production.