Senior Engineer Interview Guide

System Design Interview Questions

Assess architecture judgment with prompts for scalability, data modeling, reliability, observability, and trade-off communication.

System design interviews should reveal how candidates clarify requirements, model constraints, choose trade-offs, and communicate uncertainty. The best answers are not diagrams alone; they include operational thinking.

Book A Demo
Section 1

Scalability

Evaluate capacity estimation, bottleneck discovery, partitioning, caching, and graceful degradation.

1. Design a notification system that sends email, SMS, and push notifications to millions of users.

Interview prompt

Strong signal

Looks for queues, rate limits, retries, provider failover, user preferences, idempotency, and observability.

Follow-up probes

  • How do you handle provider outages?
  • How do you prevent duplicate sends?

Red flags

  • Sends synchronously from the API.
  • No backpressure or retry strategy.
Section 2

Data Modeling

Test schema design, access patterns, consistency requirements, and migration strategy.

1. Design the data model for a collaborative interview scorecard used by multiple reviewers.

Interview prompt

Strong signal

Considers permissions, audit history, concurrent edits, score versions, comments, and reporting queries.

Follow-up probes

  • What needs to be immutable?
  • How would you support analytics?

Red flags

  • Models only the happy path.
  • No audit trail or ownership model.
Section 3

Reliability

Assess failure modes, recovery, SLOs, redundancy, and incident readiness.

1. A video interview upload pipeline must not lose candidate submissions. Design it.

Interview prompt

Strong signal

Mentions durable storage, resumable uploads, job queues, retries, checksums, status tracking, and alerts.

Follow-up probes

  • What happens when processing fails?
  • How do candidates know upload status?

Red flags

  • Stores only in memory.
  • No retry, checksum, or status model.
Section 4

Caching

Check cache invalidation, TTLs, consistency, and performance trade-offs.

1. A public pricing page is slow and changes a few times a month. How would you cache it?

Interview prompt

Strong signal

Understands CDN caching, ISR/static generation, invalidation, stale-while-revalidate, and correctness boundaries.

Follow-up probes

  • What should never be cached?
  • How do you verify cache behavior?

Red flags

  • Caches personalized data.
  • No invalidation plan.
Section 5

Observability

Evaluate whether candidates design systems that can be operated after launch.

1. What dashboards and alerts would you add for an AI interview service?

Interview prompt

Strong signal

Looks for latency, completion rate, error rate, queue depth, provider failures, transcript quality, and business metrics.

Follow-up probes

  • Which alert pages someone at 3am?
  • How do you separate model issues from product issues?

Red flags

  • Only logs errors.
  • No user-impact metrics.

Hire better with structured interviews

Access our full question library and automate your evaluation workflow today.