Openai·Backend Engineer·Onsite - System Design / Architecture·Senior

SeniorPrefer not to say

Jun 2026

Summary

System design round at OpenAI for a backend role, centered entirely on building a geo-aware points-of-interest service from scratch. Dense problem with a lot of moving parts and not a lot of hand-holding.

Questions Asked (1)

Design a Points-of-Interest service (think Google Places or Yelp) that supports storing location data with categories and metadata, querying nearby POIs by radius or bounding box, filtering by category or keyword, and ranking results by distance and rating. The system needs to handle planet-scale traffic with low latency, hot-spot regions, and near-real-time edits.

System DesignTechnical Trade-offsData Modeling

Author's notes

This is the kind of question where you think you know where to start and then realize the scope is enormous about five minutes in.

Create a free account to read the full note

AI HintsAI Generated

Suggested Approach

Start by clarifying functional and non-functional requirements, then propose a high-level architecture that separates read and write paths. Focus on geospatial indexing (e.g., geohash or S2) for efficient nearby queries, and discuss trade-offs between consistency, latency, and scalability. Conclude with how to handle hot-spots and near-real-time updates.

Pro tip: Demonstrate awareness of real-world constraints by discussing how to handle skewed workloads (e.g., popular cities) through techniques like cell-based partitioning and caching, and mention the importance of idempotent writes for near-real-time edits.

1. Clarify Requirements

Ask about scale (QPS, data size), latency targets, consistency needs, and query patterns (radius vs. bounding box, filters, ranking).

2. High-Level Architecture

Propose a microservices-based system with separate services for writes (POI management) and reads (query serving), backed by a distributed database and caching layer.

3. Data Modeling & Geospatial Indexing

Design a schema for POIs with location, categories, and metadata. Choose a geospatial index like geohash or S2 for efficient spatial queries, and discuss how to support filtering and ranking.

4. Scalability & Hot-Spot Handling

Explain partitioning strategies (e.g., by geohash prefix) to distribute load, and techniques like read replicas, caching, and load shedding to handle hot-spot regions.

5. Near-Real-Time Updates & Consistency

Describe how to propagate edits quickly (e.g., via change data capture or write-through cache) and discuss trade-offs between consistency and latency.

Key Points to Mention

Geospatial indexing techniques (geohash, S2, R-tree) and their trade-offs
Partitioning and sharding strategies to handle planet-scale data and hot-spots
Caching strategies (e.g., Redis) for low-latency reads and reducing database load
Ranking algorithm combining distance and rating, possibly with machine learning
Handling near-real-time updates with eventual consistency and idempotent writes
Monitoring and auto-scaling to adapt to traffic patterns

AI-generated suggestions, not part of the candidate's original notes. May be inaccurate — verify before relying on them.