The hard way: get very good at C++, and focus on low-latency concerns like cache performance and code optimization. After a few years, you'll be able to make a lot of money writing trading algorithms. You won't be working on the macroscopically attractive problems-- distributed systems, functional programming languages, machine learning-- but low-latency trading has its own challenges. Some people really enjoy it. Others hate it. It's worth trying out, though. If you excel at this, you could be in the $500,000/year range by age 30, and $1 million+ is possible with a hell of a lot of luck. You will, however, have some job volatility. Your employer might vanish after a bad trade, or you might get fired for weird political reasons that have nothing to do with you. You'll win in the long term, though. The boring way: join a large software company like Google or Microsoft right out of school. Don't make any enemies, and don't take any risks. Be reliable, and develop (over years) the advantage of institutional knowledge: you know how things are done At <Company>. Focus on getting more visible projects more than on playing the promotion game or getting interesting work. If you pick the right projects, the promotions come naturally. Since you're young, you have a little more mobility than you would have at an older age, because you're more likely to get the benefit of the doubt if you have a mediocre political-success-review (err, I mean "performance" review) history and want to swing to a better project. After 8 years of grinding, you should have a title like "Staff Software Engineer" and you'll probably be right around $250,000 per year, all-in.

There are twoways to get to that level. One is boring, and the other is hard, but they both have high probabilities of success in getting you to the $250k level by age 30. Both will require a lot of focus, and you'll be tempted away by more attractive offerings like positions at startups. Don't let this happen if your goal is just to reach the $250k level.