American Express Data Scientist candidates report project-led technical interviews, SQL and ML discussion, business cases, and behavioral conversations.
$130K
Avg. Base Comp
$148K
Avg. Total Comp
2-4 rounds
Typical Rounds
3-4 weeks
Process Length
American Express Data Scientist interviews reported by candidates place substantial weight on explaining past work clearly. Expect to walk through projects or internships in detail: why you selected a model, how you processed data, what tradeoffs you made, and how the result connected to a business problem. Several candidates describe follow-up questions on tree-based models, including random forest, XGBoost, bagging versus boosting, overfitting, and regularization.
SQL and applied reasoning recur alongside model knowledge. Candidates report SQL exercises involving ranking and query optimization, as well as Python or dataset-analysis discussion. Business-oriented questions have covered credit-risk projects, customer purchasing power or credit limits, and the American Express business model. A strong response should structure the problem, state the factors and metrics you would examine, and explain how the analysis would inform a decision.
Behavioral and communication evaluation also appears throughout the accounts. Be ready for questions about feedback, setbacks, collaboration, conflict, mistakes, and why the role interests you, while keeping project explanations understandable to non-specialists. Probability questions, puzzles, simulations, guesstimates, and case studies also appear in some reports, so practice explaining a step-by-step approach when the prompt is unfamiliar. Reported formats vary: some candidates describe two interviews close together, while others report three or four rounds involving technical and manager-level conversations.
Synthesized from 14 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 American Express 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 American Express
Select the 2nd highest salary in the engineering department
| Question | |
|---|---|
| Find the First Non-Repeating Character in a String | |
| Bagging vs Boosting | |
| Variable Error | |
| P-value to a Layman | |
| Z and t-Tests | |
| New Partner Card | |
| Precision and Recall | |
| Xgboost vs Random Forest | |
| Hurdles In Data Projects | |
| Holiday Airport Traffic Surge | |
| Success Measurement | |
| Lasso vs Ridge | |
| Assumptions of Linear Regression | |
| 85% vs 82% | |
| User Event Data Pipeline | |
| Offer Matching API Design | |
| Loan Model | |
| Overfit Avoidance | |
| Optimize Model Performance | |
| Rider Acquisition Target | |
| Text Editor With OOP | |
| Fixed-Length Arrays: Deletion | |
| Acquisition Threshold | |
| Decision Tree Evaluation | |
| Different Card | |
| Client Solution Pushback | |
| Stakeholder Communication | |
| Random Forest from Scratch | |
| Why Do You Want to Work With Us |
Synthesized from candidate reports. Individual experiences may vary.
Candidates sometimes report a short recruiter call or an introductory interview focused on background, interest in the role, and behavioral examples. Others began directly with project or internship discussion, so the opening format may vary by team.
Candidates report detailed follow-ups on prior projects, data processing, model selection, tree-based methods, overfitting, class imbalance, and classification metrics. Expect to explain reasoning in plain language rather than simply list techniques.
Reported technical work includes SQL walkthroughs using joins, grouping, and window functions, plus Python-based dataset analysis. Some candidates also encountered probability questions or puzzles, so clear step-by-step reasoning may matter.
Candidates report cases involving approval-rate declines, fraud, credit limits, customer purchasing power, and merchant spend, along with guesstimates. Typically, the task is to structure an investigation, identify useful factors, and connect an approach to business impact.
Later conversations may assess business understanding, communication, collaboration, feedback, setbacks, and how you handle mistakes. Candidates also report manager or panel formats, but these were not consistent across every account.
