Lyft Data Scientist interviews reported here emphasize product metrics, experimentation, probability, and diagnosing changes in business or advertising performance.
$166K
Avg. Base Comp
$245K
Avg. Total Comp
2 rounds
Typical Rounds
1 week
Process Length
Lyft Data Scientist candidates in these reports should prepare to turn statistical knowledge into clear product decisions. One candidate described a recruiter screen followed by a 45-minute phone interview with a data scientist. That conversation combined motivation and prior-experience prompts with a metric-diagnosis exercise, a business-oriented binomial problem, and probability questions. The same report also named churn, coupon behavior, driver conversion experiments, distributions, p-values, and Bayesian inference.
A second candidate received an offer after an interview focused on product analytics, experimentation, and machine learning for advertising. Their questions covered designing an A/B test for an ad-ranking change, choosing success metrics and significance criteria, investigating a CTR decline after launch, and monitoring advertising-platform health beyond CTR. Follow-up discussion included confounding, cohort segmentation, and explaining correlation versus causation to a product manager. They were also asked to describe a data science project that influenced a product decision.
Practice explaining the diagnostic path, not just the final metric: validate inputs, segment results where relevant, weigh experiment evidence, and communicate a recommendation to cross-functional partners. The available reports are limited, so treat the reported sequence as one candidate’s experience rather than a universal format.
Synthesized from 2 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 Lyft process.
The interview focused on product analytics, experimentation, and machine learning for advertising.
There were several questions covering:
The interviewer also asked follow-up questions on handling confounding factors, segmenting results by user cohorts, and communicating recommendations to cross-functional stakeholders.
For one of the questions about investigating a CTR decline, I described verifying the data pipeline and logging first, then segmenting by user cohort, device type, and region. In my case, the new ranking model underperformed for a specific segment because it overweighted a feature that did not generalize well. I confirmed with A/B test results, recommended adjusting feature weights rather than a full rollback, and CTR recovered after retraining.
I received an offer.
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.
| Question | |
|---|---|
| Experiment Validity | |
| 500 Cards | |
| Button AB Test | |
| Raining in Seattle | |
| Impression Reach | |
| Lazy Raters | |
| WAU vs Open Rates | |
| Network Experiment Design | |
| Random Bucketing | |
| Revenue Retention | |
| P-value to a Layman | |
| Fair Coin | |
| Found Item | |
| Ride Coupon | |
| Expected Tests | |
| Estimated Rounds | |
| Three Zebras | |
| One Million Rides | |
| Uber User Journey | |
| Median Probability | |
| Secret Wins | |
| Lyft Ops Dashboard | |
| Cancellation Fees | |
| Success Measurement | |
| Biased five out of six | |
| Testing Price Increase | |
| CTR by Age | |
| Rider Discount | |
| HHT or HTT |
Synthesized from candidate reports. Individual experiences may vary.
One candidate reports a recruiter screen before the first technical interview. They encountered standard questions about interest in the role and relevant experience, so prepare a concise account that connects your past work to product-facing data science.
One candidate described a 45-minute phone screen with a data scientist. The discussion moved into diagnosing unusual metric behavior, a business-oriented binomial problem, and probability or quantitative fundamentals rather than coding-heavy work.
Another candidate reported questions on ad-ranking A/B-test design, CTR-drop diagnosis, platform-health metrics, causality, and a past project that influenced a product decision. Follow-ups may probe confounding, cohort segmentation, and stakeholder communication.
