What is system design, and where do you start?
Shows whether you frame requirements before selecting components.
Study guide
What Is System Design?Practice library
System Design Interview Questions
A good system design interview answer is a sequence of decisions, not a tour of familiar services. Choose a question, read the linked guide, then practice clarifying scope, naming tradeoffs, and explaining what changes when the system grows or fails.
How to use this question bank
- 1Pick one prompt at your current level and write down the constraints you need to clarify.
- 2Study the linked guide for the request path, data boundary, failure behavior, and tradeoffs.
- 3Repeat the answer aloud, then add one 10× scale change and one dependency failure.
Question category
Beginner
Questions that test whether you can turn an ambiguous prompt into a clear first design.
When would you scale up versus add more instances?
Tests your grasp of state, failure domains, and operational cost.
Study guide
Horizontal vs Vertical ScalingHow does a load balancer improve a service?
Opens a discussion about routing, health checks, and draining.
Study guide
Load Balancing in System DesignQuestion category
Intermediate
Questions that move from a single service to traffic, data, and product constraints.
How would you find and remove the first scalability bottleneck?
Rewards measurement-led evolution instead of premature complexity.
Study guide
Scalability in System DesignHow would you protect an API from bursts and abusive clients?
Tests identity, fairness, and distributed counter tradeoffs.
Study guide
Rate Limiting in System DesignHow do you choose SQL or NoSQL for a new feature?
Reveals whether data invariants and access patterns drive the decision.
Study guide
SQL vs NoSQL in System DesignQuestion category
Advanced
Questions where partial failure, data ownership, and operational recovery matter.
How do you shard a growing database without creating a hot partition?
Tests routing, resharding, and uneven workload reasoning.
Study guide
Database Sharding in System DesignWhen is eventual consistency acceptable to users?
Separates a slogan from a product-level correctness decision.
Study guide
Strong vs Eventual ConsistencyHow would you design a distributed cache service?
Brings together partitioning, invalidation, and degraded reads.
Study guide
Design a Distributed CacheQuestion category
Senior Engineer
Questions that require a complete service boundary and a credible operating model.
How would you design a notification system across email, push, and SMS?
Tests delivery state, preference handling, retries, and provider failure.
Study guide
Design a Notification SystemHow would you design payments without duplicate charges?
Tests idempotency, durable records, and reconciliation.
Study guide
Design a Payment SystemHow would you build a durable job scheduler?
Tests time semantics, leases, retries, and observability.
Study guide
Design a Job SchedulerQuestion category
Staff Engineer
Questions that expose how you manage organizational boundaries, migration, and sustained scale.
How would a social feed evolve for celebrity traffic and ranking changes?
Tests a hybrid architecture and evolution path, not just fanout vocabulary.
Study guide
Design a Social Media FeedHow would you separate metadata and data paths in file storage?
Tests durable control planes, lifecycle policy, and recovery tradeoffs.
Study guide
Design a File Storage SystemHow do you make a service degrade without hiding a larger incident?
Tests isolation, priorities, and observability under failure.
Study guide
Fault Tolerance in Distributed SystemsQuestion category
Principal Engineer
Questions that ask for system-wide invariants, business risk management, and multi-team decision making.
How would you protect scarce inventory during a high-demand ticket sale?
Tests authority boundaries, fairness, and reconciliation across systems.
Study guide
Design a Ticket Booking SystemHow do you reason about failures that are partial and hard to observe?
Tests coordination assumptions, timeouts, and operational strategy.
Study guide
Distributed Systems FundamentalsWhat does a 99.99% availability goal leave out?
Tests the distinction between an uptime number and dependable customer outcomes.
Study guide
Availability vs ReliabilityQuestion category
Distributed Systems
Coordination and communication questions where no node sees the entire truth at once.
What must a service decide during a network partition?
Tests explicit consistency and availability choices.
Study guide
Strong vs Eventual ConsistencyHow would you keep a chat experience useful during a partial outage?
Connects ordering, offline delivery, and connection management.
Study guide
Design a Chat SystemWhen should work move through a queue rather than a request path?
Tests acknowledgement, retry, and consumer-lag reasoning.
Study guide
Queue vs StreamQuestion category
Backend
Interface and service-boundary questions for practical backend interview loops.
How do REST, GraphQL, and gRPC change an API boundary?
Tests contract design rather than protocol preference.
Study guide
REST vs GraphQLHow would you create a redirect service that stays fast under read-heavy traffic?
Tests identifiers, caching, lifecycle management, and abuse controls.
Study guide
Design a URL ShortenerWhen should a client use HTTP versus a persistent socket?
Tests delivery expectations and connection cost.
Study guide
HTTP vs WebSocketsQuestion category
Cloud
Questions about operational boundaries, elasticity, and recovery in managed environments.
How does horizontal scaling change deployment and failure strategy?
Tests statelessness, coordination, and health-based routing.
Study guide
Horizontal vs Vertical ScalingHow do you protect durable files against a regional or processing failure?
Tests durability claims, lifecycle workflows, and repair.
Study guide
Design a File Storage SystemWhich failure should a resilient service isolate first?
Tests blast-radius thinking instead of generic retry advice.
Study guide
Fault Tolerance in Distributed SystemsQuestion category
Data
Questions about data models, replication, and high-volume ownership boundaries.
What is the difference between sharding and replication?
Tests whether you separate scale-out from redundancy.
Study guide
Sharding vs ReplicationHow do you select a shard key that can survive growth?
Tests distribution, routing, and resharding plans.
Study guide
Database Sharding in System DesignWhich queries and invariants lead you to a relational design?
Tests data-first technology selection.
Study guide
SQL vs NoSQL in System DesignQuestion category
Reliability
Questions about designing for uncertain dependencies and recovering customer trust.
How do availability and reliability differ in a customer-visible service?
Tests whether a dashboard metric is tied to a meaningful outcome.
Study guide
Availability vs ReliabilityWhat failure modes should a payment workflow make visible?
Tests reconciliation and uncertain external results.
Study guide
Design a Payment SystemHow do queues help and hurt failure recovery?
Tests lag, retries, poison messages, and overload handling.
Study guide
Queue vs StreamBefore you start
The habits interviewers can actually observe
- Clarify who the user is and which outcome cannot be wrong.
- State scale assumptions before using them to justify a component.
- Keep one source of truth for each important decision.
- Describe both the normal request path and the failure path.
- Name the tradeoff you accept instead of presenting a design as universally best.
- Finish with what you would measure, test, and change at the next scale boundary.
System design interviews are common for backend, infrastructure, platform, distributed-systems, and senior engineering roles. For the broader interview process, go back to ATOFF's Interview Preparation hub.