I would use your software skills to get in the door as a software engineer, then weasel into a sciencey position. That is:
1. Find a company that has non-software jobs that you like. Look at what companies advertise on their web sites; go to a few conferences (or watch talks) to see if some talks strike you with "ah, I can and want to do this" vibe; reach out to folks you went to grad school with, etc.
2. Apply and join as a software engineer. Don't try to sit on both chairs (software and science) during the application. You can apply to a science role, but this is likely much harder after 8 years of software focus.
3. Once in, chat with folks working on what you want to work on. Talk to folks you saw give talks. Go to internal presentations, post cool plots in slack, etc. In most companies it is pretty easy to move within roles. Plus, HR is no longer in the filtering pipeline and is not tossing resumes of anyone they think does not have the chops for the position.
Good luck!
As a personal data point -- I decided, late in my math PhD, to switch from academia to the industry after completion. A few times I switched jobs I went in as a software engineer, but within a few months moved to working on things I wanted to do beyond software (algorithms for tracking, perception, signal processing, sensor fusion, etc.).
Good point, get into the right company and move to a place you like. I did that way back moving from lab to customer service (right company and right country) to get into the R&D I wanted. Easier to execute in a company that is rapidly expanding.
This is the way. I've seen this successfully done many times. The barrier to entry is much lower once you already work at a company and just about everybody needs more software engineers than they can find.