Fundbox Data Scientist interview typically runs 4 rounds: recruiter conversation, hiring manager call, team member follow-up, take-home project presentation. Timeline is about 1-2 weeks, and the process is take-home heavy with a slow follow-up.
$212K
Avg. Base Comp
$372K
Avg. Total Comp
5
Typical Rounds
2-4 weeks
Process Length
Our candidates report that Fundbox cares less about polished textbook answers and more about whether you can turn messy inputs into a defensible analytical story. The standout signal is the take-home: it wasn’t just a modeling exercise, but a test of how you think through unstructured data, from extracting industry classifications off company URLs and reviews to deciding whether new features actually explain residuals from an earlier score. That tells us the bar is not simply “build a model,” but “show judgment in feature creation, evaluation, and interpretation.”
A recurring theme is that the team seems to value creativity and rigor in equal measure. One candidate was asked to compare performance using Gini, precision-recall curves, and decile lift, then justify the evaluation split they chose. That combination suggests Fundbox wants people who can defend methodology, not just report metrics. We’ve also seen that the project scope is substantial enough to require real time investment — roughly 8 to 10 hours when done thoughtfully — so candidates who treat it like a quick homework assignment may undersell themselves.
The other pattern worth noting is the process feel: candidates didn’t describe a strong cultural read, and follow-up could be slow or unclear. In practice, that means the interview signal is concentrated heavily in the work product itself. If you’re strong at framing assumptions, explaining tradeoffs, and showing why a feature or model is useful beyond raw score improvement, you’ll align well with what Fundbox appears to reward.
Synthetized from 1 candidates 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 Fundbox process.
The take-home was the part that stood out most to me, because it was much more involved than a typical screen. I had a standard recruiter conversation first, then a call with the hiring manager about the role and my background, and after that a follow-up with a team member who walked through my experience and the project. The project itself came with some CSV files and was clearly scoped, but it still took a meaningful amount of time — I’d estimate 8 to 10 hours if you do it thoughtfully. I worked on it independently and then presented my code and findings to the team.
The assignment was pretty open-ended and felt designed to see how I think, not just whether I can crank out a model. One part asked me to determine the industry classification for each company programmatically from the company URLs and reviews, including predicting the NAICS sector. From there I had to do EDA, create features from the provided data plus my industry classification, and compare the Gini of those features individually against the provided model score. They also wanted to know whether the new features helped explain the residuals from the earlier score. The final step was to build a model to predict the target label and compare it to the provided model score using Gini, a precision-recall curve, and the percentage of positive labels by decile. I also had to justify whether I used a train/validation split when evaluating performance. The emphasis was less on a single “right” answer and more on creativity, ingenuity, and whether I understood how to handle unstructured data.
Nobody I met was rude, but I also didn’t get much of a cultural signal either way. After I sent in my code, I was told the recruiter would follow up on next steps, and then I heard nothing for 10 days and counting. That was frustrating enough that I’d be cautious about investing time here unless you’re comfortable with a fairly heavy take-home and a slow or unclear follow-up process.
Prep tip from this candidate
Be ready for a fairly open-ended take-home that combines EDA, feature engineering, and model comparison using Gini, precision-recall, and decile lift. It would also help to practice explaining how you’d classify companies from messy text/URL data and how you’d justify your validation strategy.
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 Fundbox
How would you approach taking over and assessing an existing machine learning model?
| Question | |
|---|---|
| Empty Neighborhoods | |
| 2nd Highest Salary | |
| Rolling Bank Transactions | |
| Top Three Salaries | |
| Comments Histogram | |
| Closest SAT Scores | |
| Merge Sorted Lists | |
| Cumulative Distribution | |
| Experiment Validity | |
| Button AB Test | |
| String Shift | |
| Last Transaction | |
| Alphabet Sum | |
| Prime to N | |
| Paired Products | |
| Bank Fraud Model | |
| P-value to a Layman | |
| Swipe Precision | |
| Unique Work Days | |
| Bagging vs Boosting | |
| Over-Budget Projects | |
| Third Purchase | |
| Top 3 Users | |
| Hurdles In Data Projects | |
| Find the Missing Number | |
| Scrambled Tickets | |
| Variable Error | |
| Minimum Change | |
| Maximum Profit |
Synthesized from candidate reports. Individual experiences may vary.
A standard introductory conversation with the recruiter to discuss your background, interest in the role, and basic fit for the Data Scientist position at Fundbox.
A call with the hiring manager focused on the role, your experience, and how your background aligns with the team’s needs. This stage also appears to set up the next step in the process.
A follow-up conversation with a team member who walks through your experience and the project expectations. This seems to be a lighter discussion before the main assignment.
A substantial, open-ended project using provided CSV files. The work includes programmatically classifying companies by industry from URLs and reviews, predicting NAICS sector, doing EDA, creating features, comparing Gini against a provided model score, analyzing residuals, and building a model to predict the target label with evaluation using Gini, a precision-recall curve, and decile analysis.
You present your code and findings to the team after completing the take-home. The emphasis is on how you think, your creativity, and how you handle unstructured data rather than on a single correct solution.