S07 · interview simulation

System design: ranking ancillary offers during online check-in

People, data and outputs in this scenario are simulated. Each stage is recorded when you commit; the reveal comes after.

5stages
45 minabout
0stages committed

The brief

A system-design interview (simulated). A synthetic airline offers extras during online check-in: seat upgrade, extra baggage, lounge access, meal, priority boarding and a few others. The interviewer (simulated) says: "Design the system that decides which offers to show each passenger, and in what order."

Facts given (all hypothetical): about 2 million check-in sessions a day; the page must render within 200 ms and the ranking service may use 50 ms at p99; 6 to 12 offers are eligible per session after inventory and policy rules; the margin per offer is known; today offers are shown in a fixed order with seat upgrade first; the team is three engineers and one data scientist.

Stage 1 of 5

You have five minutes to ask questions before sketching anything.

Which questions are most important to ask first? Select all that apply.

What the objective is (revenue per session, margin, conversion) and which guardrails must not move, such as complaints or check-in time.What the logs contain: whether outcomes exist only for offers that were shown and how positions were assigned.Which rules are hard constraints (inventory, eligibility, passengers in an open disruption) and who owns them.How stale a passenger's features may be at ranking time, and what happens if the service is slow or down.Which deep learning framework and GPU type the team wants to use for the ranking model, so that infrastructure can be ordered early.