It's good that you feel stupid. That means you realize that there's room for you to improve and that you are surrounded by people who are smarter than you.

That's much better situation than if you feel smarter than everyone around you. There's a famous saying, "If you are smarter than everyone in the room, you are probably in the wrong room."

It's totally normal to feel like that as you enter the real world.

For me, I kicked everyone ass in High School. Got humbled the first semester in college. Got humbled further in grad school. And when I entered the workforce, I felt that everyone was so amazingly good and technically stronger than me. I realize this is a good thing and that I need to keep improving.

Keep grinding!

The first few times you are a reviewers are great and wonderful experience. It gets boring quick. One out of 10 papers is a good read. The rest are badly written or unnecessarily complicated. It takes a lot of times and efforts to review a paper. I used to be quite excited to review papers. But now, I decline almost all incoming invites. Just not worth the time.

The only good argument for reviewing is community service. You basically volunteer your time to make sure the community is strong. So, if you are still submitting papers for others to review, you should be doing the reviewing yourself.

Meh, no one cares if this is on your CV.

Data/Label is usually a much harder problem than algorithm.

It always has been this way. In general, research is not about breadth, its about depth. A generalist will get you nowhere in the field.