CrackML by @ml.with.umang
Interview questions / ML System Design
ML System Design interview question

Rank nearby places beyond distance

A nearby-places product currently ranks locations only by distance. Design an ML system that ranks places across categories such as restaurants, bars, and scenic locations for each user.

hardsystem designEvidence 77/1001 source reportMeta

The 60-second answer

Use geography/category constraints for candidate generation, then personalize the final ranking. Combine distance/travel time with preferences, session intent, quality, popularity, opening hours, price, freshness, and context.

Build the answer in this order

1
Frame the problem

Use geography/category constraints for candidate generation, then personalize the final ranking.

2
Design the data path

Combine distance/travel time with preferences, session intent, quality, popularity, opening hours, price, freshness, and context.

3
Choose the modeling stack

Use retrieval for scale and a richer ranker for final ordering; evaluate by category/geography and include saves/visits/negative feedback.

4
Serve, evaluate, iterate

Handle sparse users and new places explicitly with priors, metadata/content features, and exploration.

A useful interview mental model

This is the shape of a strong answer—not a script to memorize.

01Requirements
02Data
03Model / Retrieval
04Serving
05Monitor

Senior-level signal

  • Discuss exploration for new places and geographic marketplace fairness.
  • Call out location/privacy constraints and separate retrieval recall from ranking quality.

What the interviewer is really testing

Product framing, data design, modeling choices, serving constraints, reliability, evaluation, and explicit trade-offs.

Likely follow-up questions

What changes at 10× traffic or data volume?
Which failure mode would you monitor first in production?
How would you evaluate this offline and online before rollout?

Common weak-answer patterns

  • Jumping to a model before defining the product contract.
  • Listing components without bottlenecks, metrics, or failure handling.
  • Ignoring data quality, serving latency, monitoring, and iteration.