Got all excited for an open Simulink-alternative, but it’s unfortunately a cloud-based simulation product. They seem to have the requisite ML-buzzwords included. The claims about accelerating simulation with their approach, if true, are quite impressive. Wonder if the actual users have a traditional model as a validation backup in case the ML-accelerated model shows something interesting/behaves unexpectedly. Seems that it would be sensible to somehow check the results if this is going to be used for anything remotely critical like pharmaceutical modeling as they mention in the article.