Marketplaces, gig & delivery Analytics Engineer Interview Loop
Analytics engineers here own the metrics every team argues about, like completed trips, gross order value, take rate and supply hours. The analytics engineer interview checks whether you can keep those definitions trustworthy in a business with two or three sides. There's a recruiter or hiring-manager screen and a SQL screen. Then come three to five rounds: advanced SQL, data modeling, transformation and testing, and a stakeholder conversation, plus Uber's Bar Raiser where it applies. There's no statistics round.
Companies behind this pattern
Practice by round
SQL basics at speed on trip, order and driver tables: joins, aggregations, distinct counts and date functions. Explain your join path in a sentence and move on.
Advanced SQL: window functions, ranking, rolling and cumulative metrics, deduplication, top-N with ties, and funnels across each side of the market. Then come the questions that run through it. What breaks in take rate when the source changes? What would you precompute for a city operations report, and what would you compute on demand?
End-to-end projects on marketplace data, with a concept question or two on missing and filled-in values. The projects and mini-analyses are the round: each user's third transaction, registrations by email provider, city supply and demand, and partner business modeling. The metric you produce has to match between the finance view and the operations view.
| Data project | Difficulty |
|---|---|
| Driver Lifetime Value | Easy |
| Registrations by Email Provider | Easy |
| Will New Drivers Start Driving | Medium |
| Crime Data Analysis | Medium |
| Each User's Third Transaction | Medium |