Joblift is a recruitment intelligence platform backed by a meta search engine ingesting millions of job items per day and providing the possibility to search them. We are coping with massive data loads, have an intelligent candidate matching engine using machine learning models and run around 150 microservices on our platform. In this talk, we present the road to our cloud based microservice architecture operated on Google Cloud Platform and orchestrated by Kubernetes. We will focus on the (sometimes painful) learnings and limitations we encountered and how we mitigated them. This includes, for example, how we solve the daily challenges of continuous integration and deployment using a GitOps approach and practices we use to get the best out of automatic horizontal scaling.