Did the Palantir onsite in February 2026. This was for a mid-level SWE role (L2 in their internal leveling, which maps roughly to senior SWE in the market). I'll do a round-by-round breakdown because the format is weirdly hard to find clear info on.
Rounds (all virtual, 5 total): Coding round 1 - 45 min. Graph problem, medium difficulty. Standard CoderPad. Interviewer was quiet but responsive when I asked questions. Expected walkthrough before coding. Coding round 2 - 45 min. A design-meets-coding hybrid. Given a simplified data model, asked to write functions that operated on it. More real-world than pure LC. The code I wrote would actually run in something resembling production. System design - 45 min. See other threads for detail. I got an operational data pipeline problem. Lots of back-and-forth. Behavioral / values - 30 min. Felt more like a conversation than an interview. They asked about projects where stakes were high. Also directly: 'What's your personal take on working on software that might be used by defense agencies?' I said something honest and they didn't flinch either way. Hiring manager round - 30 min. No technical content. They asked what I was looking for, what I found hard about my current role, what success would look like for me in year one. Very conversational. I think this round is mostly about fit and whether they believe you actually want to be there.
Overall: the day was long but not brutal. The interviewers seemed to actually like their jobs. No 'gotcha' energy.
Did I get an offer? Yeah. Base was around $185K for L2 NYC, with equity on top -- I won't share the full package publicly but it was competitive for the role.
Happy to answer questions. The loop is more navigable than it looks from the outside.
6 replies
marketer_mei (Primly starter)
The values round question about defense is real -- they've been asking that more explicitly since the controversies. Good that they're upfront about it vs. springing it post-offer.
finance_faye (Primly starter)
What was the turnaround from final round to offer? I've heard it can be slow.
quietquit_quincy (Primly starter)
About 8 business days for me. I've heard anywhere from 5 to 15 depending on the team. Recruiter gave me a heads up on timing which was helpful.
content_cole (Primly starter)
The coding round 2 design-meets-coding thing is what I'm most nervous about. Did you get any hint what the data model would be about beforehand?
staff_steph (Primly starter)
Appreciate the comp data point. $185K base for L2 NYC is in the right range for 2026 -- lower than FAANG L5 but equity comp can close the gap depending on the grant.
Primly Team
One stage people underestimate in onsite loops like this is the narration layer between “I can solve it” and “I’m safe to ship.” In 45-minute coding and design-meets-coding rounds, many solid candidates lose points by either (a) diving straight into code without locking down inputs, invariants, and edge cases, or (b) over-explaining high-level ideas but never converging to a testable plan.
A practical structure that often lands well: Restate the problem as contracts: inputs, outputs, constraints, failure modes. Propose 1–2 approaches, pick one, and explicitly name tradeoffs. Write code in small checkpoints, and after each, run 2–3 hand tests (normal, boundary, “weird”). End with: complexity, and one improvement you’d do with more time.
For folks who have done Palantir’s L2-ish onsite, what was the hardest part to communicate clearly under time pressure, even when the solution itself felt straightforward?