There are two critical assumptions you are making here:
1) The kind of development you do is conducive to sharing via small projects on GitHub. The first four years of my full-time programing career I was working on firm real-time signal processing algorithms designed to run on an MPI cluster of POWER or Xeon machines. My skills were focused on low-level optimizations for that environment and hardware, design and implementation of grid-distributed algorithms, and understanding how to translate radar concepts and mathematical models into code. Anything along these lines that I could have thrown together for GitHub (had it existed back then) would ave been trivial and not really demonstrated anything.
2) Everyone operates under a regime wherein they do not need permission from their current employer to release things on the side. Needing to do so adds work and can be a real impediment to even wanting to put things out for public consumption.
I don’t see why you’d bother trying to force your work into that model. If it’s not common in your industry then there is no one to compete against in GitHub stakes. Sounds like publishing a paper or being awarded a patent would have the same Effect of signalling that you’re good at what you do.
Sharing my work got me, as an Aussie working mediocre jobs, on the right person’s radar to help me land a job at Facebook. This shaved probably 10 years off my career path. That was a great return on investment and a big leveller for someone who didn’t study at a brand name US college and have an obvious path to the door of a major tech company.
I’m now in a position to hire others. I certainly don’t mind if someone doesn’t have open source code shared. It clearly not a priority for many people. But I do appreciate when people do have a good backlog of work to show as it can give a seemingly lacklustre candidate a second opportunity to shine.