Google Maps Navigation – System Design Interview Preparation

Practice designing a navigation system like Google Maps. Cover map tiles, routing algorithms, real-time traffic, and ETA calculation.

This is a full mock interview round, not a written solution to read through. The AI interviewer poses the Google Maps design problem, pushes back on your choices, and asks the follow-up questions a real interviewer would: how the system behaves as traffic grows, where the bottlenecks move, which consistency guarantees you are trading away, and what happens when a component fails. You get structured feedback and a score on requirement gathering, high level architecture, data modelling, scalability and communication at the end of every attempt.

Designing Google Maps is a recurring question in High Level Design (HLD) rounds at Google, Amazon, Meta, Microsoft, Netflix and Uber, and at high-growth product companies hiring SDE-2 and above. Working through it out loud - repeatedly, under time pressure - is what separates candidates who know the components from candidates who can defend a design.

What this question is really testing

Shortest path over a graph far too large to search directly, which means precomputation and hierarchy rather than a better Dijkstra.

Scale to establish before drawing anything

Hundreds of millions of road segments globally; a naive Dijkstra across a continent is not viable at interactive latency.
Route requests must return in a few hundred milliseconds, including traffic-aware weights.
Traffic updates arrive continuously from millions of devices, changing edge weights under live queries.

Core components

Map data store – The road graph, partitioned geographically into tiles.
Routing engine – Contraction hierarchies or similar precomputed shortcuts. The precomputation is the answer.
Traffic ingestion – Anonymised speed samples aggregated per segment per time window.
ETA service – Combines static segment cost with live and historical traffic.
Tile service – Pre-rendered map tiles per zoom level, served from a CDN.

Trade-offs to defend

Precomputation versus live search – Contraction hierarchies make continental queries milliseconds and make the graph expensive to update. Traffic changes weights constantly, so the two fight - and resolving that is the interview.
Hierarchical routing – Local roads near the endpoints, highways in the middle. It is an approximation that is almost always right and occasionally not.
ETA: historical versus live – Live traffic is accurate now and wrong for the far end of a three-hour route. Blending by time-of-arrival is the real answer.

Follow-up questions to expect

A road closes. How long until routes stop using it, and what does that cost?
Route Delhi to Mumbai in 200ms. What is precomputed and what is searched?
Two users on the same route get different ETAs 30 seconds apart. Is that a bug?