Palantir hero backdrop - Denver, Colorado skyline
LiveRefreshed weekly

The Palantir interview cheat code nobody talks about

22 verified questions from the last 14 days, the same ones Palantir is asking candidates this week. Most applicants never see them.

New questions added every 7 days.

22
Fresh questions
3.2K+
Engineers using this
$245K
Median offer at stake
Scroll
19505 Biscayne Blvd · Aventura
The Process

Five rounds, and the one with no code in it fails most people

01
30 minPhone call

Recruiter screen

Mission fit gets probed here, not only at the end. Palantir screens out people chasing prestige or pay before an engineer has spent an hour on them.

  • Have a reason that is not the money.

    The recruiter is listening for why this work rather than another data company. Vague interest in hard problems reads as a no.

  • Know Foundry from Gotham.

    They are different platforms for different customers. Mixing them up is the most common tell that you have not looked.

  • Expect a question about deployment.

    Forward deployed and product development are separate tracks with different loops. Say which one you are aiming at.

02
45 minLive coding

Technical phone screen

One or two problems in a shared editor, with about twenty minutes of behavioural questions folded into the same hour. Every round here does both.

  • The behavioural is not filler.

    Roughly a third of the round is about you, and it is scored. Prepare those stories with the same care as the code.

  • Ask before you assume.

    Prompts are left underspecified on purpose. Clarifying the shape of the input is part of what is being measured.

  • Talk while you type.

    The interviewer is following your reasoning, not reading your finished function. Silence gets recorded as uncertainty.

03
60 minWhiteboard or docNo code required

Decomposition round

A vague real world problem with no defined scope. Design a parking garage, model an outbreak, build a chess game. You are graded on how you break it apart, not on syntax.

  • Decide the scope out loud first.

    Naming the inputs, the rules and the outputs before you design is most of the score. Jumping to classes loses it.

  • Strong coders fail this one.

    There is no pattern to recognise, so the usual preparation does not transfer. Practise with a whiteboard and a timer.

  • Say what you are leaving out.

    Cutting scope on purpose reads as judgement. Trying to model everything reads as somebody who cannot prioritise.

04
60 to 90 minReal codebaseNo AI tools

Re-engineering round

A few hundred lines carrying deliberate logical defects, and often a Learning round where you read the documentation for a language you have never seen and add a feature.

  • Read the whole thing before editing.

    The bugs are logical, not syntactic. A hash map that double counts will not raise an error anywhere in the file.

  • Say what each defect costs.

    Finding it is half of it. Explaining what it breaks downstream is what separates a pass from a borderline write up.

  • Read documentation out loud.

    In the Learning round they are watching how you take in something unfamiliar. Narrating it is the whole exercise.

05
45 minThen 1 to 2 weeks

Why Palantir

The final conversation is about whether you want this specific work. Recruiters say generic answers end otherwise strong loops here more often than any technical gap.

  • Missionary, not mercenary.

    That phrase gets used internally and it is the actual bar. Compensation and prestige as motives are disqualifying answers.

  • Bring a view on the customers.

    Government and defence work is the job. Having thought about it honestly, including the parts you find hard, lands better.

  • Name a product and a use case.

    Saying what Foundry actually does for one industry beats any statement about interesting data problems.

Levels and Pay

No level codes, and the stock is priced on the day you sign

Pick a level to see how the package splits, what changes by city, and what the loop actually asks of you at that band.

LevelExperienceTotal comp
Software EngineerEntry0 to 2 yrs~$195K
Software EngineerMid2 to 5 yrs~$245K
Senior EngineerSenior5 to 9 yrs~$345K
Engineering LeadLead9+ yrs~$500K

Figures shown at New York, Palantir's largest engineering site. The Florida address above is the registered head office, not an engineering one.

Live Question Feed

What Palantir is asking right now

22 questions · new ones added every 7 days
22 Verified
DECOMPOSITION

Design a system that tracks an outbreak through a contact network over time. Start by deciding what the inputs even are.

Decomposition round
60 min
Whiteboard or doc
Reported 8 times
Open
21 more questions available

The exact questions Palantir is asking this week, tagged by round and difficulty.

Unlock All 22 Questions

Each question stays in the bank for 14 days, so everything here was asked in the last 14 days. Candidates who see them first pass at 3.2 times the normal rate.

Get Full Access
The Unfair Advantage

Why this actually works

01

Real questions, not guesses

Collected from engineers who sat the loop, then cross validated against other intel sources.

02

Fresh every week

The database updates every 7 days, so you never study stale problems.

03

Covers every round

Coding, the Decomposition round, the Re-engineering codebase, the Learning round and Why Palantir.

04

120+ companies tracked

Also interviewing at Anduril, Scale AI or Snowflake? We have their questions too.

05

Difficulty and pattern data

Tagged by difficulty, round and category, so you know where prep time goes.

06

Candidate verified only

Reported by a candidate, then cross verified against multiple data points.

3.2×
Higher pass rate
120+
Companies tracked
3.2K+
Engineers using this
FAQ

Common questions

Directly from engineers who interviewed at Palantir, then cross validated against other intel sources into a stack ranked list.

We add new questions every 7 days and retire each one after 14 days, so the bank turns over twice in the window it covers.

Some, but they are not the filter. The coding round sits at ordinary difficulty, and the rounds that decide the loop are Decomposition and Re-engineering, where pattern matching does nothing.

Not in the live rounds. The Learning round hands you real documentation instead, which is the point, and interviewers watch a shared screen throughout.

A scored conversation about whether you want this specific work. Recruiters say it ends more strong loops than any technical gap, so weight your prep accordingly.

Around $245K in New York, or roughly £135K in London. About 31 percent of that is public PLTR stock, and the grant is fixed in dollars on the day you sign.

Three to five weeks. The rounds are scheduled tightly, but the final conversation and the offer approval add one to two weeks after the loop.

No. GothamLoop covers 120+ companies including Anduril, Databricks, OpenAI, Google and Meta.

Pricing plan

Get used to winning

Monthly Access
Month by month access to the most accurate SWE interview question list on the planet
$199
Per month
3h : 45m : 23s
Lifetime Access
Unlimited access to all current and future SWE interview questions, forever.
Save 60%
$800
$329
One time
Company Interview Guides

Every company we track, each with its own loop, levels and live question feed.

$472k
Palo Alto, CA
$748k
Mountain View, CA
58%
Increase
$382k
San Francisco, CA
$561k
San Francisco, CA
47%
Increase
$347k
New York, NY
$463k
New York, NY
33%
Increase
$426k
San Jose, CA
$618k
San Francisco, CA
45%
Increase
$472k
Palo Alto, CA
$748k
Mountain View, CA
58%
Increase
$382k
San Francisco, CA
$561k
San Francisco, CA
47%
Increase
$347k
New York, NY
$463k
New York, NY
33%
Increase
$426k
San Jose, CA
$618k
San Francisco, CA
45%
Increase