I am trying to find the relevant set of keywords inside a spaceless string. An example would be:
freelancemarketingconsultant
By reading it, you can distinguish the following keywords:
freelance marketing consultant
You can see the task is not trivial, as a common confusion would be to distinguish between "free" and "lance".
Are there known (potentially NLP) techniques to extract keywords from such strings?
You can use the Viterbi algorithm to find the most likely (best) way to split the string. There's a library called wordsegment that does this in Python, and you can read more about the technique at Peter Norvig's page.
There's also a recent research project called Hashtag Master which uses neural methods to tokenize hashtags.
This isn't a common problem in English, but it's standard in languages were spaces don't split words, like Japanese. There are a variety of methods, and research continues, but Viterbi based methods generally have the best balance of speed and accuracy.