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

Design an ML system using traffic data

Design an ML system that consumes traffic data at scale. Define the prediction objective, data freshness requirements, features, model choices, online serving, evaluation, and failure handling.

hardsystem designEvidence 77/1001 source reportGoogle

The 60-second answer

Clarify the target first: ETA, congestion, incident risk, or route cost, plus freshness and latency requirements. Ingest spatiotemporal data with map matching, event-time handling, deduplication, late-data policy, and historical/context features.

Build the answer in this order

1
Frame the problem

Clarify the target first: ETA, congestion, incident risk, or route cost, plus freshness and latency requirements.

2
Design the data path

Ingest spatiotemporal data with map matching, event-time handling, deduplication, late-data policy, and historical/context features.

3
Choose the modeling stack

Start with a strong statistical/GBDT baseline; add sequence or graph models only when they show incremental value on time/geography holdouts.

4
Serve, evaluate, iterate

Serve fresh features with historical fallbacks and monitor regional drift, sensor coverage, calibration, and downstream routing impact.

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 graph dependencies, delayed or biased observations, and regional cold start.
  • Call out online/offline feature consistency and how prediction errors propagate into routing decisions.

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.