Got the rejection email three days after my onsite. No call, just an email. That part stung.
I went through the full loop: recruiter screen, two technical phone screens, then an onsite with four rounds (coding, system design, product sense, and a culture/behavioral round they call "the Palantir pitch"). Here's what I think went wrong.
The coding rounds. I can do LeetCode mediums fine but one of the onsite coding rounds required optimization I wasn't expecting. I solved the problem but my first solution was O(n^2) and I had to be prompted to improve it. In hindsight, I should have practiced talking through complexity trade-offs out loud, not just reaching the solution.
System design. I over-architected. I spent 20 minutes on distributed consensus and the interviewer kept nudging me toward "let's simplify" and I kept ignoring the hint. Big mistake. Palantir's system design round is practical, not theoretical. They want to see that you can scope.
The Palantir pitch round. This is the one I was least prepared for. Basically they ask you to make a case for a real-world use of Palantir software. I had no idea what their actual products did in depth, just surface-level stuff. I fumbled it pretty badly. If I did this over I would spend a full day reading their product docs and thinking through a specific use case I could talk about with conviction.
Behavioral. This was actually fine. I had real stories from my internships and they responded well. The STAR method holds up here.
I'm going to let it sit for six months and reapply. I know the weak spots now. Rejection stings less when it's legible.