Design a Top-K search words service
Company: LinkedIn
Role: Software Engineer
Category: System Design
Difficulty: medium
Interview Round: Technical Screen
## Scenario
Design a service that tracks user search queries and returns the **Top-K most searched words/queries**.
## Requirements
### Functional
- API to **record** a search query event (e.g., user searches "iphone 16").
- API to fetch **Top-K** queries.
- Support a time window variant:
- (a) all-time Top-K
- (b) Top-K over the last **N minutes/hours** (trending)
### Non-functional
- High write QPS (search events), low latency reads for Top-K.
- Scalable horizontally.
- Reasonable correctness (define whether exact or approximate is acceptable).
## Clarifications to address
- What is a “word” vs “query” (full string)?
- Case-folding, normalization, stop words, punctuation.
- Deduping (e.g., multiple searches by same user) or count every event.
- Multi-region and eventual consistency expectations.
Overview: The question evaluates understanding of scalable, low-latency analytics and search-event systems, including high-throughput event ingestion, Top-K aggregation across all-time and sliding windows, normalization/deduplication trade-offs, and multi-region consistency considerations.
Read the full LinkedIn Software Engineer interview experience this question came from