As security engineers, we like to think that the code we write is fairly secure. Even assuming that it is, what happens when that code is left on its own? What happens when it is changed by other engineers, and new features are added? How do we keep our code secure as it grows up?This talk will discuss writing code at organizations compromised of a mix of security engineers and other programmers. We'll discuss mistakes made over several years at a large Bay Area startup, the lessons learned from them, and how you can design your code in ways that keep it safe into the future.