If you have an idea for something novel that you're passionate, then build it. Don't worry about recruiters not understanding it or whatever, ultimately if you somewhat match the job description you will get your resume in front of a hiring manager who will likely understand it more.
If you don't have a novel idea or passion project, then build something that you would find useful. If you actually use the software you write you will be more likely to complete it and polish it into an impressive state. If it's really good and open source it might even draw attention from the larger community and wind up opening doors that no recruiter would.
Good luck. I didn't take the FAANG path but sometimes look back and wish I had done that a bit, at least in my 20s.