At Uber, software reliability is of critical importance: outages canleave riders stranded and drivers without a way to earn a living.  At the same time, Uber needs to be able to move fast in developing new features and products.  Our belief is that program analysis can play a key role in reducing the tension between these seemingly-conflicting needs.  In this talk, I will describe the philosophy of how analysis tools are deployed at Uber and how code is developed to be analyzable.  I will present some initial experience reports from deployed analyses, plans for future analyses, and some open problems that may be interesting to the broader research community.