Deep Learning (DL) is increasingly adopted to solve complex tasks such as image recognition or autonomous driving. Companies are considering the inclusion of DL components in production systems, but one of their main concerns is how to assess the quality of such systems. Mutation testing is a technique to inject artificial faults into a system, under the assumption that the capability to expose (kill) such artificial faults translate into the capability to expose also real faults. In this talk, I will provide an overview of our work in adapting the idea behind mutation testing to DL systems. First, I will cover our investigation of what is a fault in a DL system. Then, I will introduce DeepCrime which is a mutation testing tool for DL systems based on the set of real faults that we have collected. Lastly, I will go through the implemented applications of mutation testing to the various software testing tasks such as test input generation, fault localisation, and test oracle generation in the domain of autonomous systems.