Pushing back on changing requirements is also a good skill to pick up. Part of it is getting a really clear agreement on what you're actually doing - what's the real scope of work - and actively changing the agreed plan when requirements change. I also find it helpful to highlight that changing requirements will have a cost in time, etc.
There's also a an art to getting together a good MVP efficiently. You can treat lots of added requirements as shiny feature requests, and prioritize them below getting the MVP working. Having an MVP that people can try out and see it working at all goes a long way towards building belief in your work. At the same time, you want to build flexibly enough that the MVP can expand rationally to take on those added feature requests... without writing in SO MUCH generality that nothing ever gets done. Doing this well is a skill learned on the back of many failures and tedious refactors... But efficiently getting a demo+MVP is golden, even if you build up some tech debt to get it.
I'll add here that there's a ton of nuance to "pushing back on changing requirements." My rule of thumb is this: program managers should understand the expected and worst-case consequences (including baseline time, switching costs, stepping on other team members' toes, etc.) of any proposals to expand scope, and as an individual contributor it is vitally important to be vocal to ensure they understand the full dynamics.
But if and once they do understand and acknowledge the consequences, and still say a scope increase is justified, it's highly highly likely that they have more business context than you. Document the decision, of course, but fundamentally trust your team.
There's also a an art to getting together a good MVP efficiently. You can treat lots of added requirements as shiny feature requests, and prioritize them below getting the MVP working. Having an MVP that people can try out and see it working at all goes a long way towards building belief in your work. At the same time, you want to build flexibly enough that the MVP can expand rationally to take on those added feature requests... without writing in SO MUCH generality that nothing ever gets done. Doing this well is a skill learned on the back of many failures and tedious refactors... But efficiently getting a demo+MVP is golden, even if you build up some tech debt to get it.