[fsloth]

Your original header and the content of your question do not match. I've been a software engineer for 10 years, enjoy my job considerably and been commended for it, but am not working for the big four or a cool startup.

I can't give you any other advice, except for my personal satisfaction the team I'm in and the domain of the problems I need to solve weights so much more than the specific company I'm working for.

For the technical interviews I've participated, I've always sucked ass. But guess what - those tests don't correlate with success at the actual work any way. They are a gate, but a bit arbitrary one at that. Yes, one needs to understand all that stuff, but can solve them at ones own pace - algorithmic puzzles usually take only a fraction of development time unless the situation is most unusual. If you can solve the problems, then from practical point of view it's sufficient.

On learning techniques, Barbara Oakley's "A mind for numbers" is an awesome book. I'm 36 and I wish I had read it 20 years ago. Slow is not bad, if correct methodology is used to verify learning. Slow can be deeper. Barbara also suggests some techniques for dealing with test anxiety - which I've not tried myself, mind you.

As per yout goal, I'm afraid I have no idea what is possible and what is not - people are complex and unpredictable. But as a senior software engineer, from the point of view of theoretical capability - if you understand the problems and can solve them at your own time, that's quite sufficient.