System Design Interview Blog

RAG system design is now a standard question in senior interview loops, and most candidates answer it as a pipeline diagram instead of a system. This guide walks through the full design with realistic interviewer dialogue at every step: why stuffing documents into the prompt fails, structure-aware chunking, the vector index memory problem, hybrid retrieval with reranking, permission-aware search, grounded generation with citations, incremental indexing and embedding model migration, and how to evaluate retrieval separately from answers. Written for senior engineers preparing for AI system design interviews at Google, Meta, Amazon, OpenAI, and Microsoft.

LLM gateway system design asks you to rate limit by tokens, not requests — and a response's token count isn't known until it finishes streaming. Covers token-aware admission with estimate-and-reconcile, multi-tenant fairness at the shared provider quota, provider fallback with circuit breakers, semantic caching trade-offs, and SSE streaming, with the interviewer dialogue senior AI infrastructure loops actually use.

Machine learning system design interviews test a different skill than coding rounds — and most candidates prepare for the wrong thing. This guide covers the 6-phase machine learning system design framework, how machine learning system design interviews differ from traditional system design, the most common questions asked at Google, Meta, Amazon, and Apple, the grading rubric interviewers actually use, common mistakes that tank strong candidates, and a 4-week study plan. For ML engineers and data scientists preparing for FAANG-level loops.

A complete iOS notes app system design guide written from an actual interview perspective. Covers local storage with Core Data and SwiftData, offline-first architecture, iCloud CloudKit sync, conflict resolution, rich text editing with TextKit 2, Core Spotlight search, and background refresh — for senior iOS engineers preparing for Apple, Google, and Notion interviews.

Master Android system design interviews with this repeatable framework. Learn to clarify requirements, design layered architectures, and handle offline-first data flow like a senior engineer.

Master monitoring system design for your next big tech interview. Covers time-series data modelling, the ingestion API, pull vs push ingestion, collector agent resilience, TSDB internals, hot/warm/cold storage tiering, downsampling, cardinality explosion, the alerting pipeline, deduplication, alert fatigue, and the exact deep-dive questions Datadog, Google, Amazon, and Meta interviewers ask.

Master distributed ID generator system design for your next big tech interview. Covers why auto-increment fails at scale, UUID trade-offs, Snowflake's 64-bit structure (timestamp, machine ID, sequence), clock skew handling, machine ID assignment, k-sortability, real-world variants from Twitter, Discord, and Instagram, and the exact deep-dive questions Google, Meta, and Amazon interviewers ask.

Master distributed task scheduler system design for your next big tech interview. Covers task submission guarantees, push vs pull worker models, leader election with etcd/ZooKeeper, heartbeat-based failure detection, at-least-once vs exactly-once execution, cron and recurring tasks, clock drift, task dependencies (DAGs), dead letter queues, and the exact deep-dive questions Amazon, Google, and Microsoft interviewers ask.

Master collaborative editor system design for your next big tech interview. Covers Operational Transformation vs CRDTs with the math explained, WebSocket architecture for concurrent editing, document storage and version history, cursor presence, offline sync, access control, and the exact deep-dive questions Google, Meta, Amazon, and Microsoft interviewers ask.

A complete Android system design guide for designing Reddit — focused on what real interviews actually ask. Covers architecture, infinite feed with Paging 3 and RemoteMediator, cursor-based pagination, Room as single source of truth, optimistic voting with WorkManager rollback, offline strategy, image loading and memory management, nested comments, push notifications, and deep linking — in a real interview conversation format for senior Android engineers.

A complete Android system design guide for designing a crypto trading app like Coinbase or Bitvavo. Covers Clean Architecture for financial apps, WebSocket channel management, real-time data pipeline with Kotlin Flow, order book rendering with throttled DiffUtil, candlestick chart architecture, Room schema for portfolio and trade history, offline resilience, price alert pipeline with WorkManager, and battery-aware subscriptions — all in a real interview conversation format for senior Android engineers.

A complete Android system design guide for designing a notes app like Google Keep or Apple Notes. Covers offline-first architecture with Room, sync conflict resolution strategies, delta sync, rich text storage with Spanned vs Markdown, full-text search with Room FTS4, WorkManager background sync, attachment handling, and multi-device consistency — all in a real interview conversation format for senior Android engineers.

Master online book review system design for your next big tech interview. Covers database schema, denormalised rating counters, Bayesian average ratings, Elasticsearch book search, activity feed fan-out for friends' reviews, the cold start problem in recommendations, collaborative filtering, content moderation, and the exact deep-dive questions Amazon, Google, and Goodreads interviewers ask.

Master Pastebin system design for your next big tech interview. Covers unique key generation approaches and their trade-offs, content storage in object storage vs database, multi-layer caching with CDN and Redis, paste expiration strategies, rate limiting, private paste security, and the deep-dive questions Google, Meta, and Amazon interviewers ask.

Master DoorDash system design for your next big tech interview. Covers the three-sided marketplace architecture, order state machine, dasher dispatch with geohashing, ETA prediction, real-time location tracking, geo-partitioned assignment service, dynamic pricing, fault tolerance when a dasher goes offline, payment idempotency, and the exact deep-dive questions DoorDash, Uber, and Amazon interviewers ask.

Master e-commerce system design for your next big tech interview. Covers microservices architecture, product catalog with Elasticsearch, Redis shopping cart, inventory management with optimistic locking, the Saga pattern for checkout, payment idempotency, flash sale handling, and the exact deep-dive questions Amazon, Shopify, and Flipkart interviewers ask.

Master real-time messaging system design for your next big tech interview. Covers WebSocket connection management, cross-server message routing with Redis Pub/Sub, Cassandra schema for messages, Snowflake IDs for ordering, delivery guarantees, group chat fan-out, presence tracking, offline sync, multi-device support, and the exact deep-dive questions Meta, Google, and Amazon interviewers ask.

Master CDN system design for your next big tech interview. Covers Anycast vs DNS-based routing, PoP architecture, multi-tier cache hierarchy, push vs pull CDNs, cache invalidation with surrogate keys and stale-while-revalidate, origin shielding, DDoS mitigation, edge compute, and the exact deep-dive questions Cloudflare, Akamai, and Amazon interviewers ask.

A complete Android system design guide for designing a photo sharing app like Instagram. Covers Clean Architecture, image loading with Glide vs Coil, multi-layer caching, RecyclerView performance, photo upload pipeline with pre-signed URLs and WorkManager, feed pagination with Paging 3, offline support, media permissions across Android versions, and memory management — all in a real interview conversation format for senior Android engineers.

A complete Android system design guide for designing Uber's mobile app — written in a conversational interview style. Covers Clean Architecture, real-time location with FusedLocationProviderClient and Foreground Service, gRPC vs WebSocket for driver tracking, map marker smoothing, trip state machine, background location permissions across Android versions, offline handling, and battery optimisation — for senior Android engineer interviews.

Master distributed metrics and logging system design for your next big tech interview. Covers push vs pull collection, Kafka ingestion pipeline, time-series database selection, the cardinality problem, hot/warm/cold storage tiering, downsampling, alerting pipelines, and the exact questions Datadog, Google, and Amazon interviewers ask.

Master Google Calendar system design for your next big tech interview. Learn database schema design, the right way to handle timezones, how recurring events actually work with RRULE, conflict detection, notification scheduling at scale, optimistic locking for concurrent edits, and how to handle free/busy queries.

A complete Android system design guide for designing a notification library. Covers library API design, FCM push pipeline, BroadcastReceiver, Android NotificationManager, persistent connections, channels, Android 13 POST_NOTIFICATIONS, in-app inbox, and key trade-offs — written in a conversational interview style for senior Android engineers.

Prepare for android system design interviews with these practice questions and a practical guide

Master Instagram and social feed system design for your next big tech interview. Learn the hybrid fan-out approach, how Instagram's real tech stack works, feed ranking with ML signals, media upload pipelines, the thundering herd problem, and how to scale to 2 billion users.

Master chat system design for your next big tech interview. Learn WebSockets, message delivery guarantees, Cassandra schema design, fan-out for group chats, presence systems, and how to scale to billions of messages a day.

Master rate limiter system design for your next big tech interview. Learn every algorithm — token bucket, sliding window, leaky bucket — plus Redis-backed distributed implementations, race condition fixes, and real-world trade-offs used by Stripe, GitHub, and Twitter.