In T341846, we built a proof-of-concept for surfacing orphan articles as recommendation to readers to increase their visibility. While this works well for some articles (example 1), the current approach times out if the article exists in many languages (example 2) due to the large number of necessary API calls to find the recommendations. This means that the current implementation is not ready to be used in practice. However, from discussions with the Web Team around annual planning indicates that the recommendations of orphans for readers would be very relevant; highlighting the need to improve the current prototype.
Therefore, in this task we want to find a better way of generating the recommendations, specifically optimizing the time needed to serve the recommendations, in order to make them more useful in practice.