Criteo Data Scientist interviews, as candidates describe them, can include a recruiter call, a data-science-flavored coding test, live problem solving, estimation cases and a take-home analysis presented to the team.
$146K
Avg. Base Comp
$162K
Avg. Total Comp
5 rounds
Typical Rounds
Not reported
Process Length
Two candidates who received Data Scientist offers at Criteo describe five rounds (one says roughly five), and both include a take-home data analysis that you present to the team. That presentation is the clearest thing to prepare for. One candidate, for whom it was the final stage, says they had to communicate key insights, justify analytical choices and field follow-up questions.
The technical content differs between the two reports. One describes a live-coding round where you explain your reasoning aloud. The same candidate had a consulting-style estimation question about the litres of paint needed to repaint Air France's fleet. The emphasis there was on structuring the problem, stating assumptions, breaking the estimate into smaller calculations and sense-checking the result. The other candidate describes a 30-minute online coding assessment with a data science slant and an in-person round with brainteasers and personality questions. They were also given a prompt to predict sales from a dataset, which called for feature engineering and model selection under time pressure. That candidate was placed on a machine learning track, which leaned heavily into the mathematical side. They advise having a solid grasp of optimization theory and maximum likelihood estimation.
Behavioral questions also appear. One candidate was asked to describe a stressful work situation, and an intern-level candidate was asked to walk through a project they were proud of. A separate intern report says the recruiter screen centered on why Criteo and the role, so keep a concise motivation answer ready. That candidate did not advance. The later stages they were told about were an online logic and numerical test and a simple technical interview.
For preparation, practice an end-to-end analysis with slides, rehearse estimation cases out loud, and review the modeling math for your track. Round counts and details vary by report, so ask your recruiter what your track will include.
Synthesized from 4 candidate reports by our editorial team.
Had an interview recently?
Share your experience. Unlock the full guide.
Real interview reports from people who went through the Criteo process.
Share your own interview experience to unlock all reports, or subscribe for full access.
Sourced from candidate reports and verified by our team.
Topics based on recent interview experiences.
Featured question at Criteo
How would you improve Google Maps?
| Question | |
|---|---|
| Drawing Balls From Bin | |
| Target Value Search | |
| Moving Window | |
| Search CTR | |
| String Palindromes | |
| Fill Rate Drop | |
| Client Solution Pushback | |
| Random Forest from Scratch | |
| Why Do You Want to Work With Us | |
| Statistically Significant Test | |
| Empty Neighborhoods | |
| 2nd Highest Salary | |
| Comments Histogram | |
| Employee Salaries | |
| Experiment Validity | |
| Subscription Overlap | |
| Merge Sorted Lists | |
| Top Three Salaries | |
| 500 Cards | |
| Button AB Test | |
| First Touch Attribution | |
| First to Six | |
| Cumulative Distribution | |
| Last Transaction | |
| Liked Pages | |
| String Shift | |
| Top 3 Users | |
| Raining in Seattle | |
| Session Difference |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report an opening call with a recruiter about background and motivation. An intern candidate was asked to introduce themselves and explain why Criteo and the role. One Data Scientist candidate says the recruiter explained the possible tracks, such as machine learning or deep learning, and was transparent throughout.
One candidate reports a 30-minute online coding assessment early in the process that leaned toward data science problems rather than pure algorithm puzzles. An intern candidate was told an online logic and numerical reasoning test would follow the screen, so the format may vary by role and level.
Candidates describe live problem solving in which you explain your reasoning as you work. One report includes predicting sales from a dataset, with feature engineering and model selection under time limits. A candidate on the machine learning track says it leaned heavily on math and advises reviewing optimization theory and maximum likelihood estimation.
One candidate reports a consulting-style market-sizing case about the paint needed for Air France's fleet, judged on structure, assumptions and sense checks. Another describes an in-person round with brainteasers and personality questions. Behavioral questions, such as describing a stressful work situation, were reported throughout the process.
Both offer-receiving Data Scientist candidates report a take-home data analysis, which one describes as the final stage. You prepare analysis and slides in advance, then present findings to the team, justify analytical choices and answer follow-up questions. One candidate called the overall process long and demanding but well structured.
