So, what is the next evolution after nanoservices? We've already seen applications being broken into smaller and smaller parts with the edition of repositories so that developers can share useful functions with others in the community.
I think we're going to see more sharing of things we can build into our applications, especially things we may not be specialists in building. I postulate that an Algorithms as a Service concept will emerge and become increasingly popular very soon. Developers who have knowledge in a particular problem domain or technology vertical will be able to share their models or the use of those models with others to achieve the same inferences.
A simple example could be a recommendation engine that predicts the likelihood of a customer purchasing a product, given their previous buying or browsing behaviors. Models could also be shared for good—think about a model that could detect a particular type of rare cancer in a photo of a number of cells. This model could be used by medical physicians to give an accurate positive or negative diagnosis, with the results of that feeding back into the training model.
This year (2019), we're seeing an explosion of new types of algorithms and new use cases. The intersection of machine learning and other technology areas has been made possible by the native integrations provided by AWS, connecting machine learning and artificial intelligence with areas such as IoT, big data, and mobile. There's no doubt that this will help drive the adoption of a new software architecture building block.
AWS has already launched its marketplace for models. AWS Marketplace for machine learning and artificial intelligence is a place where you can go to discover and buy algorithms for use in your own products or applications.