About Klarna's Hiring Philosophy

Klarna is a fintech built around payments, credit, and shopping, so its Data Scientists live close to money-critical problems: credit risk, default prediction, fraud analytics, and marketing automation. Reported assignments lean hard into these themes ("Exercise on credit scoring/default prediction", "Train a credit risk model and deploy it to some cloud service provider"), which tells you the role is genuinely applied and often end-to-end, from messy real-world data to a deployed model on AWS or another cloud.

The process is long, thorough, and famously polarizing. Candidates describe a multi-step gauntlet that "takes time" and can stretch past two months, opening with a strict logical reasoning test that eliminates people before any human conversation. Expect strong signals of Klarna's culture: their leadership principles come up repeatedly in behavioral rounds, and communication of your thinking matters as much as the final answer. Company-wide experience is genuinely split (41% positive, 45% negative), so go in prepared for both rigor and occasional slow, terse recruiter communication.

Quick Stats

* Typical process: 4 to 6 rounds, often spanning 1 to 2+ months

* Format: Mostly online (video calls, take-home assignment, remote logic tests), with some onsite for final stages historically

* Core focus: Logical reasoning, applied ML (credit risk/fraud), coding and SQL, case-study deployment, behavioral fit to Klarna principles

* Difficulty: Hard (avg 3.23/5 company-wide); the logic test cut-offs and the heavy 10+ hour ML case study are the main filters

What Klarna Looks For

* Applied ML skill on messy, imbalanced, real-world data (credit, fraud, defaults)

* Ability to build an end-to-end pipeline: EDA, modeling, and cloud deployment

* Clear communication of your reasoning, not just the final answer

* Alignment with Klarna's leadership principles and a strong culture fit

"All the rounds were not that difficult. In case you don't know the answer be frank and say you don't know rather than beating around the bush. Also let them know your thinking process as they are more interested in that rather than the answer to the solution." (Data Scientist candidate, accepted offer)

Round 1: Logical Reasoning / Cognitive Test (~15 to 30 min)

What to Expect

Almost every reported process starts here. It is typically an 18-question logical/abstract reasoning test with a tight time limit (one candidate cited "18 logical thinking ability questions" in 15 minutes) and predicting next dot patterns. Cut-offs are strict, and several candidates were rejected at this stage even when they felt they answered correctly. Later in the process, some report taking the SAME test again, this time supervised via camera. Treat it as a real filter, not a formality.

Example or Reported Questions

* "18 logical pattern questions in 15 mins"

* "The questions were about logical thinking ability"

* "Online logical test (predicting the next dot pattern)"

* "General amplitude questions, background questions"

Tips

* Practice abstract reasoning and pattern-completion tests online beforehand; multiple candidates said prep is the difference maker

* Do not "jump straight in without practicing"; one rejected candidate regretted exactly that

* Since this test bleeds straight into a recruiter chat, warm up your intro and motivation with Nora's Standard Mode so you flow smoothly from test to screen

Round 2: Recruiter / HR Screen (~30 min)

What to Expect

A short, usually friendly call with HR or a recruiter to introduce yourself, discuss your background, and align on the role and expectations. Candidates describe it as "a initial chat about 30 min" that acts as a first screen to see if your experience matches. Be ready to discuss salary expectations and relocation early. Note that some candidates found recruiter communication slow or terse, so stay proactive and keep your own timelines.

Example or Reported Questions

* "What do you currently do at your company?"

* "What do you like about your current employer's management style?"

* "Tell me your latest Data Science project?"

* "Tell me about yourself (informally)"

Tips

* Have a crisp 60-second pitch plus one recent, relevant project you can summarize clearly

* Prepare a salary range and a reason for it; one candidate was told bluntly "you are not worth that salary", so anchor with market data and stay composed

* Rehearse the "tell me about yourself" and motivation answers with Nora's Standard Mode so your pitch is tight under time pressure

Round 3: ML Case Study / Take-Home Assignment (~7 days, 10+ hours of work)

What to Expect

This is the core technical filter and the most time-consuming step. You receive a dataset (often credit risk or a classification problem) and are asked to perform EDA, build a machine learning model, and frequently deploy it to a cloud provider (AWS gets bonus points and is sometimes required via API). Expect messy, imbalanced data with many missing values. One candidate noted the task "took 10+ hours because the dataset required a lot of exploration to be understood." Briefs can be sparse, and follow-up support from HR may be slow, so document your assumptions clearly.

Example or Reported Questions

* "Train a credit risk model and deploy it to some cloud service provider."

* "Exercise on credit scoring/default prediction"

* "How to impute missing values?"

* "A difficult ML problem with AWS API implementation required"

Tips

* Build a full pipeline: EDA, handling missing values and class imbalance, a solid model (XGBoost is a popular, accepted choice), and a deployment story

* Write down every assumption; briefs are thin and data is anomalous, so make your reasoning visible and flag anything that does not add up

* Practice narrating a modeling approach out loud with Nora's Technical Mode so you can defend every choice later in the presentation round

Round 4: Skill / Technical Interview + Live Coding (~60 to 90 min)

What to Expect

You present and defend your case-study solution, then face broader technical and coding questions. Expect a live coding portion where you build an ML solution for a toy dataset in sklearn in front of the interviewer (you can pick your own algorithm and use the internet). Interviewers probe fundamentals, statistics, SQL, and your resume projects. As one accepted candidate stressed, showing your thinking process matters more than a perfect answer.

Example or Reported Questions

* "What is supervised Learning?"

* "Explain central limit theorem? What is mean? What is standard deviation?"

* "Write down a SQL query based on question asked."

* "If you can get access to all data, what kinds of variables will affect the credit model?"

Tips

* Refresh ML fundamentals, core statistics, and SQL; questions range from basics to deep ML/deep-learning detail

* Be ready to code live in sklearn and to speak every decision aloud; if stuck, ask the interviewer a clarifying question rather than going silent

* Run a mock in Nora's Technical Mode focused on ML concepts, stats, and SQL so you can explain trade-offs cleanly and think out loud on command

Round 5: Behavioral / Klarna Fit + Team Fit (~45 to 60 min)

What to Expect

A STAR-based behavioral round tied tightly to Klarna's leadership principles, sometimes split into a "Klarna fit" and a "team fit" conversation. Candidates report business-scenario and hypothetical questions (positioning Klarna vs competitors, prioritization under deadline pressure, handling a difficult teammate) alongside classic behavioral prompts. Read the Klarna principles in advance; you may be asked which matter most and which you would drop.

Example or Reported Questions

* "Which of the Klarna principles do you think is the most important and why?"

* "If you were the CEO of Klarna, what are 3 factors that make you stay awake at night?"

* "Why should Amazon use Klarna instead of Paypal?"

* "Let's say I need you to prepare 6 dashboards for tomorrow with person X, who is always doing the least. How would you approach this situation?"

Tips

* Prepare 5 to 6 STAR stories covering deadlines, conflict, ambiguity, and impact, and map each to a Klarna principle

* Have an opinion ready on Klarna vs competitors and on which principles matter most, since these come up as pointed hypotheticals

* Drill these with Nora's Behavioral Mode to tighten your STAR structure and practice the principle-based and "if you were the CEO" scenarios out loud

Frequently Asked Questions (FAQ)

1) How many rounds are there?

Typically 4 to 6. A common path is: logical reasoning test, HR screen, ML case-study take-home, a skill/technical presentation, a live coding round, and a behavioral/Klarna-fit interview. One accepted Stockholm candidate described a full "6 round interview" spanning more than two months.

2) What topics are most common?

* Applied ML on credit risk, default prediction, and fraud, including deployment to a cloud provider

* ML and statistics fundamentals, SQL, live coding in sklearn, plus Klarna-principle behavioral questions

3) How long does the process take?

Expect 1 to 2+ months. Multiple candidates said it "can take up to two months," driven by the number of rounds, the multi-day take-home, and sometimes slow recruiter communication between steps.

4) How should I prepare?

* Practice abstract/logical reasoning tests online before you start, since the strict cut-off eliminates candidates early

* Build fluency in end-to-end ML: EDA, imbalanced data, missing-value handling, a strong model, and cloud (AWS) deployment

* Review core statistics, SQL, and be able to code and narrate an sklearn solution live

* Use Nora's Technical Mode for the ML/coding rounds, Behavioral Mode for the STAR and Klarna-principle questions, Standard Mode for the recruiter screen, and Salary Negotiation Mode to anchor your pay expectations before the offer stage