I guess they're not flush with cash. From what I noticed most small German/EU companies developing complex SW projects like this, that aren't backed by well funded VC funds, don't bother with hiring dozens of juniors like other unicorns usually do since they don't have the money. They prefer to wait and only hire the "perfect candidate", usually experienced seniors withe the right skills willing to work for the salaries they can offer, which isn't much.

I also have a DeepL competitor in my EU country and their hiring practices are similar, rarely hiring, and mostly senior devs and ML PHD grads from the local university. And they're not paying insane salaries, which probably explains why they hire so rarely.

It's probably a business that's difficult to monetize successfully when Google, Microsoft, etc. are offering "free" solutions which are good enough for the casual use, and for professional use when big money is on the line, companies usually pay authorized translators(humans) anyway.