The truth is that academia in Germany is mostly precarious. There just aren’t enough jobs and with the Schuldenbremse many fundings have been cancelled. A permanent position is hard to get and doesn’t pay very well. What’s more likely to happen is that you will be stuck in a circle of non-permanent contracts including regular relocation until you might get your chance (or not).

Just to throw in some other options: you can easily convert basically anything to latex (and ultimately to Pdf) using pandoc. For instance, if you use Zettlr as your markdown editor, you can also use a citation software (eg., Zotero) and quickly invoke it using the @ character. Then, you can write your documents in Markdown and inline Latex and create Latex-powered Pdfs via pandoc. I use this approach to write scientific papers and it works pretty well.

Is there something against an approach where we only store data, like we did for the last centuries. But how the data is aggregated and presented is not defined. There could be various frontends using different (public) algorithms. This could work pretty easy with decentralised networks like the fediverse.