As others have pointed out already...I've noticed that sometimes the male voice does not always pronounce words in the way that I remember them from the time I studied Russian years ago (for example потом here). Is there a reason for this? Is this a regional, dialectal thing?
I'm Russian. The pronunciation of this word is incorrect, at least to me. The stress should be on the second syllable: https://ru.forvo.com/word/%D0%BF%D0%BE%D1%82%D0%BE%D0%BC/#ru
The reason for this is that the robot takes words out of context. The stress in Russian words often differs depending on the meaning. "ПотОм" means "then", "пОтом" is the instrumental case of the word "пот" ("sweat"). The TTS doesn't know which meaning is implied here and, it seems, chooses one randomly.
I've read elsewhere on the forum that it's possible Duo would implement a new TTS for the Russian course soon(-ish). Hopefully the new one would be better in recognizing these things.