A job in science communication after a PhD in Mathematics is not a waste at all! In fact, in my country, there is a well known example of a math PhD and science communicator (in addition to recently having become professor in science communication): Ionica Smeets (Wikipedia).
She mainly is known from co-writing a popular math-blog ('wiskundemeisjes') and more recently from presenting the yearly 'national science quiz' ('de Nationale Wetenschapsquiz') on Dutch television. As far as I know, she's never had any formal training in journalism. (but perhaps you could email her to ask if you're serious enough, she might be receptive)
Her inauguration speech (In Dutch), roughly translated is titled 'On the value of science communication' has some important paragraphs (translated and paraphrased):
When mathematicians exposed the Monty Hall problem to the general public, the responses to the correct result reported where both massive and hostile! First in the US, then when I reintroduced the problem in the Netherlands. (One reader asked whether it wasn't written by someone who has studied mathematics instead of her [she was a PhD student at the time])
She proceeds that she continued her effort to explain the correct solution to her readers.
I suggested that people try to solve the problem by repeated simulation it. One reader wrote that he spend whole night playing the game and didn't understand why, but he did understand that she was correct.
This, I think, is key. The role of science communication is both to explain and convince. If explanation fails, try to at least convince people of the scientific (here mathematical) truth.
Of course, whether people know the correct answer to a puzzle can hardly be called important. But it is a good example of what can go wrong and right in an explanation and that explaining science is far from trivial, a discipline in itself! She proceeds to show examples from medicine where the understanding of the general public is important and were completely misrepresented by journalists (not completely their fault, as I said, science communication is hard)
@DaveLRenfro mentions two more examples: Helen Joyce (Ph.D. under David Preiss) and Evelyn Lamb (Ph.D. from Rice University).