Personalization in online learning can be found in several domains, such as language learning, chemistry, or math education. Going beyond these domains is rarely seen. Due to fixed learning progressions, learners enrolled must complete the entire course to succeed. Learners’ prior knowledge is often not considered. This can lead to disengagement as learners are not challenged appropriately if task difficulty levels do not align with their skills. To allow a flexible course entry, we introduce a placement test to estimate learners’ competency levels employing the item response theory and implement a computerized competency-based adaptive placement test in real-world online courses of the KI-Campus. We illustrate the challenges of applying item response theory with multiple competencies in a real-world setting and demonstrate how to identify a domain-independent, optimal course entry point tailored to the learner. This application promises to optimize the learning time considering learners’ prior knowledge, which in turn, can reduce drop-out rates.