Have you ever wondered how large software companies with an engineering culture make sure they are able to deliver software over and over to production? How do you coordinate 100+ software engineers so that there are no bottlenecks and quality is not compromised? In this talk you will see how a Continuous Delivery system was implemented at Criteo, the fastest growing IT company in EMEA 2012. Before starting the project there were 160+ code repositories with dependency hell. They were being built independently and releases to production were error prone and painful. You will see the technical architecture behind a successful implementation of a Continuous Delivery system. The system was made up of a Gerrit code review tool connected to a Jenkins build pipeline, building 160 repositories with over 7M lines of code. We will explore different architectural choices such as branching system, hot fixes, sandbox and pre-production environments, and how these were developed and used by the large R&D department. Authors: Adrian Perreau de Pinninck, Manu Cupcic