I'm onboarding and might be facing a death march project soon. What do I do?

This isn't the best for an onboarding engineer, but it is an opportunity to earn trust fast. If your manager and team are reasonable (i.e. they mathematically factor in the fact that you're new and expect less, primarily from a deadline perspective), this actually is a blessing in disguise (most senior engineers have not enough scope instead of too much). My high-level pieces of advice:

1. Establish expectations - You mentioned "...I'm afraid I have to produce something meaningful within a tight deadline over the course of the next few months". Talk to your manager and clarify that: When is it due? Better yet, create a rough timeline and overall spec doc with milestones/dates. Run it by the team and ask them if it's reasonable.
2. Aggressively bring people together - Since you're new, you have the superpower of being able to ask for tons of help and face time with people. Use that to be an information sponge, and don't be afraid to ask for pair programming sessions to learn way faster.
3. Always think "I can do this!" - This sounds like an empty pep talk, but mentality really does go a long way (and it's true!). As an engineer, you are only truly limited by how fast you can type code. Let's say the project will take 2,500 lines of code. On a good day, I can write 1,000 lines of code, so the theoretical super best-case scenario is 2.5 days. Work backwards from that and break down what it takes to make that vision a reality:
   1. How can you get the requirements solidified super fast? - Talk to the PM and make them your best friend.
   2. How can you get the technical designs up super fast? - Get to know all the senior/staff engineers, make an initial doc, and cc them all for feedback. After that, hold a tech review meeting ASAP.
   3. How can you write the code super fast? - Ask tons of people to explain various parts of the codebase to you and start tinkering right now. Submit messy PRs (because you don't know any better) and address the feedback at lightning speed. Set up focus blocks to move faster.

I actually went through something similar onboarding at Robinhood. I was put on a chaotic project with no real definition or existing leading authority - It was a huge curve ball. So I filled in the gaps and became the expert on it as a newbie. I go through this experience in my case study here: [Case Study] Becoming A Tech Lead Again In Just 1 Month After Joining Robinhood From Meta

Do it! I agree it's not the ideal onboarding setup, but "ideal" is rare to find 😅 How big is your company? How understanding is your manager, and how important is this project?

If people around you recognize that the situation is not ideal, then the expectations should be calibrated accordingly. If you fail in the right way, you should be setup to move to another part of the company where you have a better chance of succeeding.

It's critical to build allies in your first few months at the company, and ask for help aggressively: Ask Questions Well.

The other part that would help answer this is.. what's the alternative? Has a team switch already been offered to you? How difficult would that be?

The failure mode of switching teams is that it appears you are 'running away' from the problem instead of tackling it directly. It'll also burn valuable time during which you could land impact. For example, at Meta, if you spend 2 months on the team switch, you put your next PSC cycle in jeopardy since a significant portion of the half is gone.