
State Farm Data Scientist candidates report a mix of one-way HireVue video screens, take-home modeling work, and technical rounds on statistics and ML fundamentals. Here is what to prepare.
$105K
Avg. Base Comp
$120K
Avg. Total Comp
3-5 rounds
Typical Rounds
Not reported
Process Length
Candidates describe the State Farm Data Scientist process as technical early, but the reports differ in format. Several mention a one-way HireVue video screen with technical prompts. One candidate was asked about multicollinearity and how to address it, and says other accounts they saw mentioned unsupervised learning and logistic regression assumptions. Another describes business-scenario prompts: handling an imbalanced target, and why a top random forest feature might not be significant in a logistic regression.
One report describes a four-day take-home. It asked for data cleaning, a logistic regression, a second model (the candidate chose XGBoost), predictions on a test set, and an executive summary. That candidate says the later technical discussion was expected to probe imputation, leakage, and how to explain AUC to a business audience. Another candidate describes a take-home with a performance threshold before the video interview.
Live rounds seem to reward clear explanations. One candidate describes a statistics-heavy round covering R-squared, correlation, confusion matrices and error metrics, followed by a discussion of how State Farm operates and where data science fits. Another was faulted for staying at a high level when asked about gradient boosted logistic regression. A senior-title report spent most of a 45-minute round on univariate analysis, imputation and outlier handling.
All six reports here ended in no offer, so treat these as examples of what was asked, not as a fixed process. Practice explaining each technique's mechanics, assumptions and tradeoffs in plain language, and practice answering on camera within a time limit.
Synthesized from 6 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 State Farm 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 State Farm
What are the assumptions of linear regression?
| Question | |
|---|---|
| P-value to a Layman | |
| Rectangle Overlap | |
| Bias vs. Variance Tradeoff | |
| Insurance Leads | |
| Multicollinearity in Regression | |
| Double Card Value | |
| Area Under the ROC Curve | |
| Hurdles In Data Projects | |
| Xgboost vs Random Forest | |
| Data Preparation for Imbalanced Data | |
| Bias Variance Tradeoff | |
| Digitizing Student Test Scores | |
| Your Strengths and Weaknesses | |
| Different Card | |
| 2nd Highest Salary | |
| Employee Salaries | |
| Top Three Salaries | |
| Experiment Validity | |
| Random SQL Sample | |
| Bagging vs Boosting | |
| Fair Coin | |
| Lasso vs Ridge | |
| Duplicate Rows | |
| Size of Joins | |
| Always Excited Users | |
| Get Top N Frequent Words | |
| Total Spent on Products | |
| Boarding Times Bias | |
| Booking Regression |
Synthesized from candidate reports. Individual experiences may vary.
Candidates report different openings. One was invited to a HireVue the day after applying, and another describes an eligibility and overview conversation about what data science is and how they could contribute. A third reports a 30-minute phone screen. Be ready for a clear background walkthrough and a short answer on why you want the role.
Several candidates describe recording answers with little or no live interaction. Reported prompts include introducing yourself, why you fit, multicollinearity, handling imbalanced data, and explaining a key model feature to a business partner. One candidate also heard of questions on unsupervised learning and logistic regression assumptions. Practice concise spoken answers with a short thinking window.
Two candidates mention a take-home. One received it after a phone screen and describes four days to clean simulated data, build a logistic regression and a second model, predict on a test set, and write an executive summary. Another had to clear a performance threshold before the video interview. Document your imputation, leakage checks and model choices, since the exercise may be discussed later.
Candidates report live technical rounds focused on fundamentals: R-squared, correlation, confusion matrices, error metrics, logit assumptions, imputation, outliers, and gradient boosted logistic regression. Interviewers may press on how and why a method works, so prepare step-by-step explanations with tradeoffs, not just names.
One candidate reports a final technical discussion on how State Farm operates and where data science creates value, plus a behavioral question about completing a task with minimal instruction. Another was told later rounds may include hiring manager, senior data scientist and peer conversations. Prepare examples of working independently and of where data science fits the business.