Semantic segmentation models based on deep learning techniques have been successfully applied across a wide range of domains. However, their adoption can be challenging for non-expert users due to several factors. These include the need to experiment with different algorithms implemented across heterogeneous libraries, the use of different annotation formats, or the inconsistency in evaluation metrics used across different tools. In this work, we present the first steps towards building an AutoML framework that facilitates the construction of segmentation models implemented across multiple libraries. The framework supports users throughout the entire pipeline including the analysis and split of datasets of images, the training of the models, and their evaluation. In addition, the framework also provides the necessary tools to apply Zero-shot Active Learning by reducing the annotation effort. Thus, this framework will help to lower the entry barrier for applying state-of-the-art segmentation techniques.