You seem to be putting "the right way" on some kind of pedestal. There is no one right way. There are right ways to do software in a lab, there are right ways to do software in a startup. You've learned those two. You lack exposure to one of the right ways to do software in a megacorp that optimizes for making developers fungible resources and maximizing shareholder value. It doesn't make you somehow less of a developer not being able to do the third. If you want to chase a high TC then by all means, jump in and learn it. But, that's the only reason. Learning how to develop for a megacorp won't enrich your life, enlighten you or make you feel anymore like you've made it as a developer.
Also, the grass is greener won't stop. Once you get to the megacorp you will wax poetic about your academia and startup days. You'll tell those fun stories to your coworkers because you stopped accumulating fun work stories when you joined megacorp. You'll occasionally think about going back to grad school, or joining a startup again; but, you'll remember you have an RSU vest in 2 months, better stick around at least for that, can't leave money on the table...
Would it be a crazy idea to consider that deploying emotional intelligence has less obvious, higher order effects, that make up for the initial investment?
Also, the grass is greener won't stop. Once you get to the megacorp you will wax poetic about your academia and startup days. You'll tell those fun stories to your coworkers because you stopped accumulating fun work stories when you joined megacorp. You'll occasionally think about going back to grad school, or joining a startup again; but, you'll remember you have an RSU vest in 2 months, better stick around at least for that, can't leave money on the table...