Audio duration and animation timings when localizing

I’m coming across 2 separate but related challenges when localizing. 

The first challenge is slide duration. As you can see in the attached screenshot, the English-language audio has >10 seconds of silence at the end, because the Romanian version of that audio (for example) is longer. The 2 audios are “connected” and the longer audio is driving slide duration.

The second challenge is that because the Romanian version is longer, the animations that are timed to the English version (e.g. ovals 1/2/3 in screenshot) are not timed well for Romanian.

I think that one solution is creating separate layers for each language: basically duplicating the base layer of every slide for each language, and displaying them using the language trigger condition. But I think this solution will end up having some limitations.

Another path I’m considering is that maybe I have to export each language as a separate course, and then manage each course separately. (really don’t want to manage 9 different courses).

Thoughts on these ideas, or any other ideas are greatly appreciated.