Data Scientist interview
at Stripe.

Tell-me-about-a-time stories on ownership, conflict, and impact — scored against this company's real values.

1. 01 Tell me about a time your analysis changed a product or business decision, and how you made sure the result was rigorous enough to act on.
2. 02 Describe a situation where you had to work backwards from a user's actual need rather than the metric you were originally asked to optimize.
3. 03 Walk me through a time you found a flaw in a widely accepted assumption or 'received wisdom' on your team and how you challenged it from first principles.
4. 04 Tell me about a time you delivered an analysis under significant ambiguity and time pressure, and what tradeoffs you made to move with urgency.

The hard screen for the craft itself — talked through out loud, not whiteboarded in your head.

1. 01 How would you design an experiment to measure whether a change to our checkout flow actually increased payment authorization rates without being confounded by fraud or retries?
2. 02 Given a table of charges with statuses, timestamps, and idempotency keys, write SQL to compute the true authorization success rate while correctly deduplicating retried payment attempts.
3. 03 How would you estimate the causal impact of rolling out a new fraud rule on legitimate-transaction conversion when you can't run a clean randomized test?
4. 04 Walk me through how you'd build a model to forecast daily payment volume that must remain reproducible and auditable for finance reconciliation.

Open-ended design of the systems this company actually runs, with the interviewer probing your tradeoffs.

1. 01 Design an analytics pipeline that produces a daily authorization-rate metric that is reproducible, reconcilable against the ledger, and trustworthy for finance.
2. 02 Design an experimentation platform for safely testing fraud and checkout changes across millions of merchants with guardrail metrics.
3. 03 Design a data model and lineage approach for tracking every state transition of a payment from creation to settlement.
4. 04 Design a system to detect and alert on anomalies in key payment metrics like dispute rate or false-positive fraud blocks.