Serverless computing (aka Function-as-a-Service) provides developers with a serverless deployment and operations model relieving them from the need to worry about complex infrastructural and operational tasks and allowing them to focus on developing value adding code. It guarantees an optimal utilization at any scale where developers do not have to pay for resources idling around but for code really running. I.e. it is about a model where code only runs on-demand, on a per-request basis. OpenWhisk is an open-source FaaS platform, driven by IBM and partners like Adobe, that allows you to execute code in response to an event and that is ideally suited to a wide range of scenarios including cognitive, data, IoT, microservices, and mobile workloads. During this talk, we will talk about our journey through the world of serverless computing, the core concepts, the key value proposition and differentiators, typical usage scenarios, and the underlying programming model of serverless computing in general and OpenWhisk in particular. We will also discuss some of our latest additions and illustrate how to benefit by “going” serverless with OpenWhisk by exploring some real-world customer usecases with a focus on how serverless architectures can be exploited for cognitive applications such as multimedia (images & videos) analysis & object detection, and conversational applications. Using these usecases we will explain how OpenWhisk works and why it is the ideal platform for these emerging workloads. We will conclude our session with some basic demos.