Apache UIMA is a great framework for integrating multiple annotators in a NLP pipeline. This type of approaches have fall out of preference in the academic community with the emergence of end-to-end (E2E) deep learning (DL) systems.
In industrial systems, although, there is value in extending pipelines containing DL components.
For DL components, a project that is also under the Apache Software Foundation is the MXNet project. Integrating neural networks in the ONNX format in a way similar to the ClearTK annotators would be very useful.