While the new advances in AI research are exciting, startups and companies are still struggling with the adoption of machine learning. The operationalisation of machine learning remains a key challenge for startups and bigger companies. Fortunately, in recent years there has been a lot of development and movement around the practices, methodologies and tooling that are addressing the needs to build reliable machine learning systems. This new field is called Machine Learning Operations (in short MLOps). It’s a multidisciplinary field that combines data science, software engineering and devops. In other words MLOps bridges the gap to apply machine learning in the real world. Through this talk we explore MLOps and unveil the challenges it solves and why it makes sense for companies to jump on the MLOps train.