Please fix Swahili's English
A large number of English translations are embarrassingly bad or wrong. From others' comments, it seems that these have not been fixed for a year, which is also embarrassing.
It would help if you gave examples. But yes, you're right. The course contributors are currently MIA.
Lately, the updates are unsigned. It used to tell us that Emilian Mbassa or one of the other contributors did it, but now it just says "Duolingo". They also seem to only be correcting the English at the moment, so my gut feeling is that it's someone in the Duolingo team who maybe doesn't know Swahili well but can at least work on the English where it's clear. There are only a few Swahili sentences that actually have mistakes, but there are lots and lots where there are several possible answers and not all are accepted.
What Ben said is completely accurate. I know several people who have applied to be contributors and have heard nothing back. Duolingo seems to have decided this course isn't as important as adding pictures to the app and things like that.
Sentence accepted e-mails for all courses are now unsigned. When they were signed Emilian (or Rehema less commonly), they came in spurts. They still come in spurts. If Duolingo had just assigned somebody to "clean up," I'd expect a more continuous trickle of e-mails.
Once Duolingo picks the one or two people to head up the team, it leaves it up to them whether they want to incorporate additional contributors or not. For now it appears this course was born in a Peace Corps office, and they're keeping it there. Ukrainian and Guarani have transitioned out of that model, but it took a while. (For Guarani it was demonstrably Duolingo central that dropped the ball, letting the course sit entirely contributor-less for something like a year.) At least the Swahili contributors have remained and are observably working. Would it be good if they onboarded some more people? Yes, I would certainly think so, in particular people with knowledge of other Duolingo trees that could apply that familiarity to help with a reconceptualization of the tree (fewer sentence fragments!) looking toward a version 2.0 and beyond.