Palantir · Phone screen → Onsite
In processPalantir FDSE New Grad Interviews: 5 Experiences (Part 1 of 2)
Palantir FDSE new grad interviews, 2025–2026: a 20-minute screen, then learning and decomposition rounds; Why Palantir came up in three of four reports.
At a glance
- Process: a 20-minute recruiter or engineer screen (01, 03) → a learning round where you read and use unfamiliar code (01, 03) → a decomposition round on an open design problem (03) → hiring manager. One referral's recruiter said a 45-minute coding round, a virtual interview and an onsite would follow (04).
- Asked recently: Why Palantir in 3 of the 4 reports (02, 03, 04); resume and project talk in all 4; learning an unfamiliar query API in 2 (01, 03).
- Difficulty: the 2025 New York candidate called the coding task no harder than LeetCode easy, but did not pass the onsite (03); the 2026 candidate got stuck at first in learning and needed hints (01).
- Practise first: Map and Accumulate Learning Queries, Relational Learning Queries, Organism Knowledge System and Why do you want to work at Palantir?
Coverage: 5 Palantir Forward Deployed Software Engineer (FDSE / FDE) new grad candidates (reports 01–05). Four are current, reported April 2025 to October 2026, in New York where a place is given; three came through referrals. One September 2023 report sits in History.
In this series
- Part 1 (current): the newer reports, 01–04, plus H05: screens, a map-and-accumulation learning round and an exoplanet decomposition.
- Part 2: six older reports, 2018–2023: algorithm, SQL and REST OAs, an employee treasure hunt and the typosquatting OA.
Other Palantir cohorts: Internship · Experienced hires
What came up
Palantir's FDSE loop tests how fast you learn and how you break down a vague problem, more than classic algorithms. Counts are out of the 4 current reports.
| Topic | Reports | Last seen |
|---|---|---|
| Resume, projects or background | 4 of 4 | Oct 2026 |
| Why Palantir and why this role | 3 of 4 | Oct 2025 |
| Learning and using an unfamiliar API | 2 of 4 | Oct 2026 |
| Open-ended design (decomposition) | 1 of 4 | Apr 2025 |
Questions we found
All dates are report months.
Learning rounds (coding)
| Question | Reports | Last seen |
|---|---|---|
| Map and Accumulate Learning Queries | 1 | Oct 2026 |
| Relational Learning Queries | 1 | Apr 2025 |
Decomposition (system design)
| Question | Reports | Last seen |
|---|---|---|
| Shared-Interest Connection App | 1 | Apr 2025 |
| Organism Knowledge System | 1 | Apr 2025 |
Behavioral and recruiter (BQ)
| Question | Reports | Last seen |
|---|---|---|
| Why Palantir? | 3 | Oct 2025 |
| Introduce yourself | 2 | Apr 2025 |
| Why this role? | 1 | Oct 2025 |
| Why did you do a PhD? | 1 | Apr 2025 |
| Are you open to travel? | 1 | Apr 2025 |
| Do you have other offers? | 1 | Apr 2025 |
4 detailed candidate timelines
01: A 45-minute learning round with map and accumulation
Report date: October 28, 2026 · Role: FDSE, Delta · Application: Referral · Result: Rejected after learning
Screen (20 minutes)
- About a week after a friend's referral, the candidate expected a recruiter call. An FDSE ran it instead: recruiter-style talk, resume and project questions, and one or two technical questions. Tying their background to Delta's work seemed to help.
- Both this screen and the later learning round were moved at the last minute.
Learning (45 minutes)
- The invite came about half a week later. Learning comes first and must be passed before decomposition.
- About 15 minutes of routine behavioral talk, then 30 minutes on the task.
- The candidate read code for a SQL-like class API, explained it back, then used its operations: one function mapped values into a data structure, another worked like a loop or accumulation. They got stuck at first, took small hints and finished two parts with three questions. Practice: Map and Accumulate Learning Queries, which spells out the data and goals the report left out.
- A rejection followed, so there was no decomposition round. The candidate wondered whether speed hurt them; no feedback said so.
02: A short HR call about motivation
Report date: October 16, 2025 · Role: FDSE · Application: Referral · Round: HR · Result: Passed
- The recruiter went through the resume and focused on why Palantir and why this role.
- The candidate suggested knowing the products and their strengths before the call. No technical task was described.
03: From an employee-interest app to an exoplanet
Report date: April 14, 2025 · Role: Forward Deployed Software Engineer, new grad · Location: New York onsite · Result: Rejected before the hiring manager
Phone screen
- Applied in late February. In early March, a 20-minute screen: introduction, resume projects and why Palantir.
Video interview
- About a week later: ten minutes on the resume, then ten to twenty on a Shared-Interest Connection App that helps people with similar interests meet.
- A short coding task followed, unusual for LeetCode and no harder than easy. It was not named.
Onsite decomposition (one hour)
- You are the first person on an exoplanet, alone. Design a system that records every living organism there.
- The interviewer kept widening it: back on Earth, how would you use the first trip's data to build a better biological information system, and how would you train the next explorers? Practice: Organism Knowledge System.
Onsite learning (one hour)
- The interviewer first explained relational databases, including one-to-many and many-to-many links.
- The candidate then used a Python-object query builder, SQL-like but not an ORM, to implement the requested functions. Practice: Relational Learning Queries. The individual queries were not shared.
- The two rounds ran back to back, about four hours at the office in total, including time with other candidates; this group was among the first back to in-person onsites that year.
- The candidate did not reach the hiring manager.
04: HR questions for a PhD candidate
Report date: April 4, 2025 · Role: written as "FDSW"; the candidate asked how FDSE works with other teams · Location: New York · Result: Passed HR
- A friend's referral, and HR scheduled the call quickly.
- The recruiter asked: tell me about yourself; why did you pursue a PhD?; why Palantir?; are you open to travel (yes); do you have other offers?
- The candidate asked how FDSE works with product and other teams.
- HR said the next steps were a 45-minute coding round, a virtual interview and an onsite. The candidate had not taken them yet.
Patterns across the reports
- Why Palantir gets real time. Reports 02, 03 and 04 all spent real time on it; in 02 it was the main topic.
- Learning means reading code you have never seen. Both 01 and 03 had to understand a query-style API and then use it, but 01 had about 30 minutes and 03 a full hour.
- Referrals move fast. Reports 01, 02 and 04 came through friends; 01 was contacted in about a week and 04 heard back quickly.
Common questions
What is the Palantir learning round?
You get unfamiliar code or a concept, explain it, then use it: a SQL-like class API with map and accumulate steps in 01, and a relational query builder in 03.
What is the Palantir decomposition interview?
An open-ended design problem that keeps growing: recording every organism on an exoplanet, then using that data on Earth and training the next team (03).
How long is the Palantir FDSE phone screen?
About 20 minutes in 01 and 03; in 01 an FDSE ran it rather than a recruiter.
History (before Nov 2023)
This report is older than three years. It counts toward this part but stays out of the summary and tables above.
H05: A new grad loop with undisclosed technical tasks
Report date: September 26, 2023 · Role: FDSE, full-time new grad · Result: Rejected
- Applied August 1 after Palantir webinars and most of a day on the essays; no internship experience. An invite came the next day.
- HR, September 1 (moved from August 31): their studies and move from sociology to computer science, about 20 minutes on why Palantir, and sponsorship (needed).
- September 6: onsite invite with compensation expectations, prep material and product information, plus an optional decomposition prep session.
- September 21: one hour of learning, then one hour of decomposition. Interviewers challenged choices and asked the candidate to defend them; systems, frontend and backend, SQL and operating systems came up. The tasks were not shared. The candidate felt they missed a follow-up.
- September 25: rejection citing limited experience, with an invitation to reapply.
Keep practicing
- Map and Accumulate Learning Queries: report 01's learning round.
- Relational Learning Queries, Shared-Interest Connection App and Organism Knowledge System: report 03.
- Why Palantir?, Why FDSE? and Why a PhD?: reports 02, 03 and 04.
- Palantir coding bank, system design bank and behavioral bank: all Palantir questions by type (sign-in needed).
- Palantir interview guide and Palantir hiring insights: stages, practice and recent hiring signals.
- Experience feed: interview reports from other companies and roles.
In this series: Part 1 (current) · Part 2
Sources: Anywhere useful interview info can be found.
Comments
0No comments yet — be the first to share what you know.