Katkat

Topic

System Design interview questions

1311 System Design questions from 230 companies, with hints, runnable code and solution walkthroughs.

Questions1311

60 easy819 medium432 hard

Your progress0/1311

System Design interview questions by the numbers

Last updated

Katkat has 1,311 System Design interview questions from 230 companies; the most come from OpenAI (8%), Meta (5%) and Amazon (4%). By difficulty: 5% easy, 62% medium and 33% hard. The most common round is Onsite (55%). Frequent concepts include System Design, Distributed Systems, Scalability and Caching.

Companies (230)

Difficulty

  • Easy5% · 60
  • Medium62% · 819
  • Hard33% · 432

Interview rounds

  • Onsite55% · 721
  • Technical Screen39% · 517
  • Online Assessment3% · 38
  • Other2% · 21
  • HR Screen1% · 13

Roles

  • Software Engineer91% · 1189
  • Machine Learning Engineer2% · 31
  • Data Engineer2% · 26
  • Backend Engineer2% · 25
  • Frontend Engineer1% · 19

Frequent concepts

  • System Design31% · 400
  • Distributed Systems24% · 316
  • Scalability15% · 196
  • Caching7% · 95
  • Concurrency7% · 91
  • system-design6% · 83

Showing 991–1020 of 1311 results

  • Design a typeahead autocomplete service that starts at a modest scale using an in-memory keyword trie, then scale it out without adding external services. It tests trie internals and keyword storage, read and write paths, score updates, rebuilds with hot swaps, and sharding strategy trade-offs.

    PinterestSoftware EngineerSystem DesignOnsite
  • Design a low-latency, real-time typeahead suggestion system that handles prefix search, query ranking, aggressive caching, and stale response mitigation.

    PinterestSoftware EngineerSystem DesignOnsite
  • Pinterest software-engineer onsite system-design question: design a multi-region, high-throughput distributed rate limiter targeting ~10M req/s with per-user, per-API, and global limits, burst tolerance, and an approximately-sliding window. It tests token-bucket/GCRA algorithm choice, an architecture of local in-process buckets fed by epoch-based token leases, consistent-hashing sharding with hot-key splitting, bounded-overshoot consistency, fault handling, fairness, capacity-planning math (including a 2× spike), backpressure, and observability.

    PinterestSoftware EngineerSystem DesignOnsite
  • A system design question about an ad click aggregator whose queries cover the last year rather than the last few minutes and which must take its click data from an existing logging API. It tests requirement clarification, checkpointed ingestion and backfill, deduplication, rollup storage, and correcting historical counts.

    PinterestSoftware EngineerSystem DesignOnsite
  • This question tests the ability to design a high-throughput, fault-tolerant system for tracking ad interaction events, focusing on scalability, exactly-once processing, and analytical query performance.

    PinterestSoftware EngineerSystem DesignOnsite
  • This question evaluates a candidate's competency in designing scalable, durable, and highly available blob/object storage systems, encompassing distributed systems architecture, data modeling for buckets and objects, metadata management, partitioning and routing, fault tolerance strategies, consistency models, large-object transfer patterns, security controls, and operational concerns like monitoring and capacity planning. Commonly asked in the System Design domain to assess practical application of architectural patterns and trade-offs—horizontal scalability, durability versus latency, consistency guarantees, and operational resilience—it primarily tests practical system-design skills grounded in conceptual understanding of distributed systems.

    PinterestSoftware EngineerSystem DesignOnsite
  • Design a high-scale merchant data enrichment platform that ingests, reconciles, and maintains up-to-date business attributes (addresses, social metrics, wiki metadata) across heterogeneous third-party sources with strong provenance and fault tolerance.

    PinterestSoftware EngineerSystem DesignOnsite
  • A Kubernetes scheduling exercise for a senior infrastructure role: four applications request scheduling at the same time, and the candidate must ask the questions that settle capacity, priority, placement and coordination, then commit to a plan. A second part designs a scheduler plugin that uses signals reported by nodes.

    PinterestSoftware EngineerSystem DesignTechnical Screen
  • This question evaluates a candidate's ability to design large-scale conversational AI systems, covering competencies in distributed system architecture, real-time streaming, stateful multi-turn session management, safety controls, and integration with external tools.

    PlaidSoftware EngineerSystem DesignTechnical Screen
  • This question evaluates a candidate's competency in designing geo-distributed, globally consistent rate limiting systems that enforce per-API-key and additional constraints while balancing low-latency request paths and regional failure tolerance.

    PlaidSoftware EngineerSystem DesignTechnical Screen
  • Design a secure, embeddable Pay by Bank frontend widget and supporting backend infrastructure that enables direct bank checkout via an isolated iframe and asynchronous payment processing.

    PlaidSoftware EngineerSystem DesignOnsite
  • This question evaluates proficiency in designing production data orchestration, data quality validation, and observability frameworks for both batch and streaming pipelines, covering dependency management, schema and integrity checks, and operational handling of anomalous records.

    Point72Data EngineerSystem DesignTechnical Screen
  • Design analytics for a small independent e-commerce store, with the emphasis on how behavioral and purchase data is gathered in the first place. It tests event modeling, browser versus server-side capture, visitor identity across checkout, proportionate storage and funnel reporting, privacy, and data-quality reconciliation.

    PosthogSoftware EngineerSystem DesignTechnical Screen
  • Control LLM workflow quotas and 429 retries with shared token budgets, durable checkpoints, and explicit handling of uncertain requests and duplicate billing.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Design notification ingestion and delivery for a million internal producers, separating service-count scale from throughput, bursts, and fanout.

    Qualified HealthSoftware EngineerSystem DesignOnline Assessment
  • Isolate tenants and sessions across RAG retrieval, caches, prompts, and tools, with server-enforced authorization and layered prompt-injection defenses.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Design a 200-document RAG session pipeline with asynchronous ingestion, corpus readiness, latency budgeting, scoped caching, and quota-aware serving.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Design 24-hour document deletion, seven-day chat history, fast sidebar reads, and cold audits while resolving conflicts over derived content and physical erasure.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Recover DAG workflows through crashes, duplicate attempts, external side effects, and multi-day approvals using durable state and explicit exactly-once limits.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Scale mixed I/O and LLM workflows with separate worker budgets, tenant fairness, tiered state storage, and explicit reconciliation of conflicting workload figures.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Stream provisional LLM output to clients while durably validating and committing complete structured results before downstream DAG tasks execute.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Design SSE or WebSocket streaming, end-to-end stop handling, and safe Agent retries while distinguishing local cancellation from provider billing and tool effects.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • Validate editable workflow DAGs, reason about cycle-check latency, and enforce durable retry and expansion limits that prevent unbounded runtime loops.

    Qualified HealthSoftware EngineerSystem DesignOnsite
  • This question evaluates expertise in concurrent programming, lock-free MPMC queue usage, producer–consumer asynchronous logging, variadic formatting APIs, and robustness in shutdown/flush semantics and error handling.

    Qube Research & TechnologiesSoftware EngineerSystem DesignTechnical Screen
  • This question evaluates a candidate's ability to design scalable, fault-tolerant real-time analytics systems for maintaining a sliding 60-second visitor count, covering competencies in event ingestion pipelines, deduplication, late-event handling, storage selection, scaling, and monitoring.

    RampSoftware EngineerSystem DesignTechnical Screen
  • Design a high-throughput, low-latency room availability and reservation service that prevents double-booking while maintaining fast read times under heavy search traffic.

    RampSoftware EngineerSystem DesignTechnical Screen
  • Tests candidate's ability to design a real-time, low-latency financial metrics dashboard that evolves from a simple relational database setup to a horizontally scalable, stream-aggregated architecture resilient to duplicate events and counter drift.

    RampSoftware EngineerSystem DesignOnsite
  • This question tests your ability to design a robust, distributed financial ledger system that handles high-throughput transactions, ensures strict consistency, and supports the complex operation of merging two distinct user accounts.

    RampSoftware EngineerSystem DesignOnline Assessment
  • This interview question evaluates requirements, scale assumptions, API/data design, architecture, trade-offs, failure modes, and rollout in a realistic interview setting. A strong answer for Design an in-memory cloud storage system states assumptions, handles edge cases, explains trade-offs, and shows how to validate the result clearly.

    RampSoftware EngineerSystem DesignOnline Assessment
  • A senior system design prompt reported from a RapidSOS loop: design the product catalog for a multi-store grocery delivery marketplace like Instacart. It tests modeling shared products versus per-store prices and availability, ingesting messy retailer feeds, store-scoped browse and search, caching, freshness and checkout consistency.

    RapidSOSSoftware EngineerSystem DesignOther

More topics