I think if you want to have high impact, agency, and autonomy as an IC late in your career you should work on your math skills with an eye towards a research sc

I think if you want to have high impact, agency, and autonomy as an IC late in your career you should work on your math skills with an eye towards a research scientist / core algorithm designer position. The other option is to be the founding engineer at a startup, but that's probably only good in your 40s if you want a divorce