Design a Food-Item Review System

Quick Overview

Explore the data model, consistency, moderation, aggregation, and scale trade-offs behind a food-item review platform.

Design a Food-Item Review System

Company: DoorDash

Role: Software Engineer

Category: System Design

Difficulty: hard

Interview Round: Onsite

# Design a Food-Item Review System Design a system where customers can submit and read reviews for individual food items. Explain the product contract, write and read paths, aggregate ratings, moderation boundaries, and scaling choices. ### Constraints & Assumptions - A review belongs to a specific food item and author identity. - Edits and deletions must not silently corrupt displayed aggregates. - The design should separate confirmed requirements from assumptions about ordering, anonymity, and moderation. ### Clarifying Questions to Ask - Can one author review an item more than once? - Must reviews be tied to a completed order? - How fresh must the displayed average and count be after an edit? ```hint Separate source and projection Treat individual reviews as the record of truth and the displayed rating summary as a derived projection. ``` ### What a Strong Answer Covers - Core APIs and data model for items, reviews, authorship, and aggregate summaries. - Consistency choices for create, edit, delete, and rating recomputation. - Pagination, sorting, caching, abuse controls, and moderation or audit history. - Failure handling, observability, and a path to rebuild derived aggregates. ### Follow-up Questions 1. How would verified-purchase labels change the model? 2. How would you rank reviews without making the aggregate rating unstable?

Quick Answer: Explore the data model, consistency, moderation, aggregation, and scale trade-offs behind a food-item review platform.

|Home/System Design/DoorDash
DoorDash logo
DoorDash
Sep 1, 2026
hardSoftware EngineerOnsiteSystem Design
2
0

Design a Food-Item Review System

Design a system where customers can submit and read reviews for individual food items. Explain the product contract, write and read paths, aggregate ratings, moderation boundaries, and scaling choices.

Constraints & Assumptions

  • A review belongs to a specific food item and author identity.
  • Edits and deletions must not silently corrupt displayed aggregates.
  • The design should separate confirmed requirements from assumptions about ordering, anonymity, and moderation.

Clarifying Questions to Ask Guidance

  • Can one author review an item more than once?
  • Must reviews be tied to a completed order?
  • How fresh must the displayed average and count be after an edit?

What a Strong Answer Covers Guidance

  • Core APIs and data model for items, reviews, authorship, and aggregate summaries.
  • Consistency choices for create, edit, delete, and rating recomputation.
  • Pagination, sorting, caching, abuse controls, and moderation or audit history.
  • Failure handling, observability, and a path to rebuild derived aggregates.

Follow-up Questions Guidance

  1. How would verified-purchase labels change the model?
  2. How would you rank reviews without making the aggregate rating unstable?

Submit Your Answer to Earn 20XP

Sign in to leave a comment

Loading comments...