FAANGs have plenty of hard problems - most of them are involved in hardware, robotics, distributed systems, machine-learning, huge data pipelines, compilers, et

FAANGs have plenty of hard problems - most of them are involved in hardware, robotics, distributed systems, machine-learning, huge data pipelines, compilers, etc. However, you usually will not get to work on those directly as a new grad, and for several years later. The reason for that is simply that they are hard problems, and big corporations are risk averse, and so they want people who have already made lots of mistakes working on the hard problems so that they will hopefully make fewer mistakes. At Google at least, the "hard" interesting technical work usually starts around L6 or high L5, which will take you 5-10 years to achieve. (The exceptions are some truly exceptional undergrads that got into the particular problem domain as a high-school student, and so already have 5-10 years of experience at 23.) If you want to short-circuit this process, your best bet is often to pick your domain, start looking into it as a hobby, and use this to get a job at a startup in the field that's desperate for people. Because they're desperate, they're often willing to take the risk of putting junior engineers on hard problems, because, well, they have no other alternatives. And a FAANG background makes you look great in these cases. Then once you've got a couple years of experience in the domain, come back to a FAANG, but this time do it with a few years of relevant experience, which qualifies you for the hard problems. Make sure you leave on a high note; this makes you eligible for rehire and often lets you skip the re-interview process entirely.