Design a delivery cost aggregator
Overview
This question evaluates skills in designing a concurrent in-memory aggregation component, including system design, thread-safety, time and currency precision, and live-metrics computation.
Rippling
Aug 7, 2025, 12:00 AM
Software Engineer
Technical Screen
System Design
0
0
In-Memory Driver Payments Aggregator
Goal
Design and implement an in-memory component that tracks driver payments and exposes one metric: the total cost of all deliveries (for a live dashboard).
APIs to Provide
- addDriver(driverId: int, usdHourlyRate: double)
- recordDelivery(driverId: int, startTime, endTime)
- getTotalCost(): double
Requirements and Assumptions
- Drivers have different hourly rates.
- Drivers may have overlapping deliveries and are paid for each delivery independently.
- Each delivery is recorded immediately after completion and lasts no more than 3 hours.
- Time fields must support at least 1-second precision.
- getTotalCost returns the aggregate cost across all drivers for a live dashboard.
- All inputs are valid; exact output formatting is not required.
Clarifications to Make Explicit
- Time representation: Use UTC Unix epoch seconds (int64/long) for startTime and endTime to avoid timezone/DST issues and meet the "≥ 1-second precision" requirement.
- Currency arithmetic: Avoid accumulating money in double to prevent rounding drift on the live dashboard.
- Thread-safety: Assume the component may be called concurrently (dashboard + event ingestion). Make the implementation thread-safe.
Solution
Login required
Show