J'ai postulé en ligne. Le processus a pris 1 jour. J'ai passé un entretien chez Palantir Technologies en avr. 2012
Entretien
Having heard how hard Palantir interviews were, I answered the phone expecting to fail.
They gave an introduction and asked what I would change about their product.
Then went straight into algorithms.
Asked how to find the median in a list. Then a more efficient solution than my answer. They helped/guided me through how to find the optimal answer, but I got there in the end.
Next asked how to keep track of minimum number in a stack without losing O(1) complexity. I answered and once again, was asked to improve efficiency, which I did.
I asked how I could improve in future interviews, he said that I did well and that there wasn't really anything I could have done better and that thinking a loud was good. Hung up the phone thinking I did very well, got sent rejection letter two weeks later. Not really sure what they wanted, perhaps you need optimal answers from the start.
Follow up to question 1: If inserting a 1000 numbers then finding the median a 1000 times, and you had to choose between either a logn insert algorithm or a logn median look up, which would be better?
J'ai passé un entretien chez Palantir Technologies (Miami, FL) en juin 2026
Entretien
Started with a recruiter screen where the whole point is just checking if you actually care about their mission and the real-world impact of their software, rather than just wanting a cool tech job. After that was a 90 minute hackerrank OA that felt more like an implementation mini-project with SQL and Python instead of abstract algorithms.
The onsite was a 4-round loop chosen from decomp, re-engineering, learning, coding, and sys design. Decomp is the most important one - they give you a super vague prompt like designing a chess game or tracking a disease from scratch, and you have to map out the inputs and logic out loud. Re-engineering gives you around 1000 lines of code with a very subtle logical bug to fix, and the learning round drops you into a random API with barely any documentation to see how fast you pick it up lol. Coding was standard LC mediums but they squeeze a 20-minute behavioral chat right into the middle of it, and sys design was heavy on data governance and fault tolerance. The final chat with the hiring manager is pretty intense too ngl. They will actually make you redo parts of the onsite you struggled with. For prep, don't just mindlessly grind LeetCode. Practice reading other people's code fast and structuring ambiguous problems. I got a really good Palantir coach on Prepfully who helped a lot to catch my blind spots and get a reality check before the actual loop. Overall, not very easy though
Questions d'entretien [1]
Question 1
A payment processing module has a race condition that produces incorrect totals under concurrent writes. Walk through how you would identify the root cause and propose a fix.
J'ai passé un entretien chez Palantir Technologies (New York, NY)
Entretien
Great interview process - 1. Recruiter call 2. Leetcode style technical 3. Scoping style (decomp) interview 4. Frontend coding 5. Another scoping (decomp round).
Interviewers were fun and engaging, and I felt challenged in a positive way.
Questions d'entretien [1]
Question 1
Why do you want to work here?
What are you looking for in your next role.
J'ai passé un entretien chez Palantir Technologies
Entretien
Recruiter flaked me 3 times and this was always during the time of the interview. I would join the interview meeting and the recruiter would say ahh sorry I got a conflict, next time.