Yelp is the dominant US local search and review platform — 200M+ reviews of local businesses. The interview emphasizes search relevance, local data, fraud detection (fake reviews), and the unique two-sided marketplace dynamics of consumer + business.
Process
Recruiter screen → 60-minute coding phone (DSA medium) → onsite virtual: 2 coding, 1 system design, 1 craft deep-dive, 1 behavioral. Cycle: 3–4 weeks.
What they actually ask
- Design a local search ranking system (geographic + relevance + quality). Be ready to combine a spatial index (S2 cells or geohash) for proximity, text-match relevance against business name and category, and quality signals like rating, review count, and recency into a single score. Interviewers push on how you weight and tune those signals, and how results stay fresh as new reviews come in.
- Design fake review detection. Focus on the signals — account age, review velocity, device/IP clustering, and near-duplicate text — and whether you’d use rules, a trained classifier, or both. They probe the precision/recall tradeoff, since flagging a real reviewer costs as much as missing a fake one.
- Design business profile management with claim verification. Walk through the ownership-claim flow (verify by automated phone call or a mailed postcode code), deduping listings for the same storefront, and moderating conflicting edits. Expect follow-ups on how you resolve two parties claiming the same business.
- Coding: medium DSA, often with graph or geo framing. Practice grid and graph traversal (BFS/DFS), interval merging, and bucketing points into geographic cells — the core patterns tend to show up wrapped in a maps or reviews scenario. A common example is grouping nearby locations or finding the k nearest businesses to a point.
- Behavioral: customer focus, ownership, working with mature platform. Prepare STAR stories about a consumer-versus-business tradeoff, shipping in older code you didn’t write, and owning an ambiguous problem end to end. They want signal that you can move a stable, well-tested platform without breaking it.
Levels and comp (2026)
- SE II: $150K–$185K total
- Senior SE: $215K–$285K
- Staff: $310K–$410K
- Principal: $430K–$570K
Prep priorities
- Be fluent in Python (legacy ML/data) or Go (newer services). Know one deeply — Python still runs much of the legacy ML and data tooling, while newer services lean on Go. Be ready to write clean, tested code and talk through concurrency in whichever you choose.
- Understand search and ranking systems. Review how an inverted index, candidate retrieval, and a scoring/ranking layer fit together, and how you’d A/B test a ranking change. Reasoning about relevance-versus-freshness tradeoffs matters more here than any single algorithm.
- Brush up on geographic indexing (S2, geohash) and local-relevance signals. Understand how S2 cells and geohashes turn lat/long into indexable buckets that make radius and nearest-neighbor queries fast. Know the local signals that move ranking — distance, category match, rating, and review recency.
Frequently Asked Questions
Is Yelp remote-friendly?
Yes. Fully remote in the US since 2020.
How does Yelp compare to Google Maps or Tripadvisor?
Yelp is local-business specific. Google Maps wins on map experience. Tripadvisor specializes in travel. Yelp comp is mid-tier; below FAANG.
What is the engineering culture?
Mature, stable, distributed-first. Strong testing and deploy discipline. Mission has stabilized post-IPO.
Similar company guides
Prepping for Yelp? Put it to work:
