did you consider tier 2, semiconductor suppliers qualcomm, nxp or go tier 3 - arm? I would say these will allow more technical depth than tier 1s, but will remove you from the end customer product, and some people find that unsatisfying.

customer facing companies with engineering culture also come to mind, apple, google etc. I assume you are in Europe? you have Bosch Research, Fraunhofer or Dolby in US, this might overlap with your stated experience.

I dont have experience in quants or trading but I heard they can go pretty deep and hard.

I get the impression you are not technically challenged at your work. Not an uncommon dynamic when working for a company for which software is a means to an end.

In order to dig deep into one area, you might need to take on a "hobby" project that forces you to deep dive in a chosen area. Looking at your current areas of experience, I would choose either GPU or DSP programming. By being CPU adjacent areas, specialist skills are more likely help your career.

If you are at a profitable physical goods company for 8 years and still employed there, I think you're doing pretty well.

That said, if you want to switch jobs (in general) you mostly need 1) the exact skills that that job requires, which are listed on job listings usually, so read those and b) probably you need to know someone who works there already to put your profile forward so you can get the foot in the door for an interview and rise above the 2000 other similarly qualified applicants. There are only so many industries and companies, it's not insurmountable to research.

What are you interested in doing at this "new job"? C, C++, python, OpenCL, GPU, DSP, and some docker again, but in more depth? Then you'd want to look for jobs that have those as requirements. Do you know people at any of the companies with those job listings? Is that realistic? Maybe, IDK, you can look and find out though.

Like I said, you are maybe luckier than you know.

I tell you my experience. I have >10 YOE. In a niche. And everyone whats to hire for the EXACT KNOWLEDGE THEY NEED not what some smart person has and can do once hired. So I cram for like 5 different domains for 5 different interviews, I tell you want, it's silly.

Today's hype/value cycle is in AI.

If you want a 30-day plan to upskill to deliver that to companies, see the post at https://news.ycombinator.com/item?id=48218162 and the tweet linked in the OP. The outline some kind of 30 day plan to learn those skills. If that sounds like something for you, maybe consider it.

If you just want generic depth in your favorite technical languages/topics, maybe start contributing to the open source projects in those languages that you like or make use of. Probably a pretty good way to get depth. But better to solve real problems for money, maybe. Because nothing ain't free anymore.

Honestly, best advice is this book IMO, "The Science of Getting Rich" by Wallace D. Wattles (by now old enough to be in the public domain):

  Read: https://www.gutenberg.org/files/59844/59844-h/59844-h.htm

  Listen (3 versions): https://librivox.org/author/848?primary_key=848

It sounds like a motivational self-help or spiritual book, but if you can get past that, the main idea is to not throw away our progress, but instead:

  1) always build on what you've built already in terms of expertise and experience and ability to add value, and

  2) always give people more value (at work) than they expect or think they deserve

I think (2) is your ticket to adding depth and (1) is your way to finding that path.

One other advice, I find every morning it helps to think: how great it is this thing I'm working on today because it helps people who need _____ (that thing). It makes it more exciting. Maybe you don't need this one, but I thought I'd mention it.