Design Maps Nearby

"Design Google Maps" sounds enormous, but it almost always narrows to one feature: nearby search. Type "pizza," get the places around you, ranked sensibly, in a blink. Google reaches for it because the core problem is a query shape the coding rounds never touch. A database finds a row by key in microseconds. Ask it for "everything within two kilometers of me" and the same database falls over. The design exists to fix that.

Anchor on the numbers first. Google Maps tracks roughly 200 million places, and each one is about a kilobyte of metadata: name, coordinates, category, rating. Multiply those and the entire planet's place data is around 200 GB. That fits on a handful of machines. So this is a latency and query-shape problem, and storage is the one thing we will never worry about. Searches run on the order of a billion a day. Call it 10,000 a second, with sharp peaks as time zones wake up. Edits to places trickle in a few hundred times a second at most. Reads win by a few hundred to one, and the reads are the product.

The Google angle

Maps is a Google flagship, so the person across the doc has probably worked near the real thing and will steer with confidence. That cuts two ways. The format is still a shared Google Doc with no canvas, so rehearse describing grid cells (the map squares this whole design hangs on) and indexes in plain sentences. And brace for the escalation. Once your static-places design holds up, many Google interviewers flip it to "now find nearby friends," where the points move. Deep dive five takes that head on, and it's the turn that catches candidates flat.