System Design Interview – For Architects

Social Media & Content Platforms

Applied "design X" prompts on social media & content platforms. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q1. Design a news feed system for a social network with 500 million daily active users, where celebrities can have over 100 million followers.

Q2. Design a system to display real-time 'like' and 'share' counts on posts viewed by millions of users simultaneously.

Q3. Design a content moderation pipeline that can review and act on user-generated posts, images, and videos within seconds of upload.

Q4. Design a system for showing users 'who viewed your profile' with near real-time updates, at the scale of a professional networking platform.

Q5. Design a system to detect and remove fake accounts and bot-driven engagement at the scale of a major social platform.

Q6. Design a system for a photo-sharing app that must serve billions of images with multiple resolutions, globally, with sub-100ms load times.

Q7. Design a system to support 'stories' style ephemeral content that automatically expires after 24 hours, viewed by millions of users per day.

Q8. Design a hashtag trending system that must identify and rank the top trending topics across a platform in near real-time.

Q9. Design a direct messaging system that must support both one-on-one and large group chats with thousands of participants.

Q10. Design a system for showing personalized 'people you may know' suggestions at social-network scale.

Q11. Design a system to handle a viral post that suddenly receives 10 million views within an hour, without degrading the rest of the platform.

Q12. Design a comment system for a platform where posts can receive millions of nested, threaded comments.

Q13. Design a system for tracking and displaying 'mutual friends' between any two users at the scale of billions of connections.

Q14. Design a system to support live commenting and reactions during a live-streamed event with millions of concurrent viewers.

Q15. Design a system to archive and later restore a user's entire account history (posts, messages, media) when they request account deletion with a 30-day grace period.

Q16. Design a system for detecting and limiting the spread of misinformation across a social platform in real time.

Q17. Design a notification system that must deliver push, email, and in-app notifications to a billion users without becoming a bottleneck during high-activity events.

Q18. Design a system for a social platform's 'memories' feature that resurfaces a user's past posts from exactly one year ago, every day, for all active users.

Q19. Design a system to support user-generated live audio rooms (like a live audio chat) supporting thousands of simultaneous listeners per room.

Q20. Design a system for detecting coordinated harassment campaigns targeting a specific user across a social platform.

Q21. Design a system to let users search across billions of historical posts by keyword, hashtag, and date range.

Q22. Design a system for a platform's 'verified badge' program that must validate identity claims for millions of applicants.

Q23. Design a system to support collaborative playlists or shared photo albums that multiple users can edit simultaneously.

Q24. Design a system for an influencer marketing platform that must track and attribute sales generated through creator content across multiple e-commerce sites.

Q25. Design a system to power a 'for you' algorithmic feed that must balance content freshness, relevance, and diversity across a billion-user platform.

Q26. Design a system for tracking and preventing spam link distribution across direct messages and comments at platform scale.

Q27. Design a system to support 'close friends' style selective content sharing, where users can post content visible only to a hand-picked subset of their followers.

Q28. Design a system for detecting when a public figure's account has likely been compromised, based on unusual posting behavior.

Q29. Design a system for a social platform's live-streaming feature that must handle sudden viewer spikes when a stream goes viral mid-broadcast.

Q30. Design a system to let advertisers target ads to specific user segments while preserving strict user privacy guarantees.

Q31. Design a system for a platform to support multi-language automatic translation of posts and comments in real time.

Q32. Design a system to track content virality and automatically identify posts likely to go viral before they actually do.

Q33. Design a system for a professional networking platform to recommend relevant job postings to users based on their profile and activity.

Q34. Design a system to handle a scenario where a platform must support both text posts and long-form video content within the same underlying feed architecture.

Q35. Design a system for detecting and handling duplicate or re-uploaded copyrighted content across a user-generated content platform.

Q36. Design a system to support 'read receipts' in a messaging platform without creating a privacy or performance problem at scale.

Q37. Design a system for a live sports platform that must show real-time score updates and stats to millions of concurrent users during a major game.

Q38. Design a system to detect when a user's account exhibits behavior consistent with account sharing or credential leaking, like simultaneous logins from geographically impossible locations.

Q39. Design a system for a content platform to support creator monetization through subscriber-only content, at the scale of millions of creators and subscribers.

Q40. Design a system to power a platform's algorithmic 'explore' page that surfaces content from accounts a user doesn't yet follow.

Q41. Design a system to support real-time collaborative story or fan-fiction writing where multiple users contribute to the same evolving narrative.

Q42. Design a system for a platform to detect and prevent engagement farming, where accounts artificially inflate metrics through bait tactics or coordinated groups.

Q43. Design a system to support a platform-wide 'trending now' page that must update within seconds as global attention shifts.

Q44. Design a system for a social platform to support scheduled posts that publish automatically at a future specified time, at scale of millions of scheduled posts.

Q45. Design a system for a platform to measure and report content reach and impressions accurately, given the complexity of what counts as a genuine 'view.'

Q46. Design a system to support a platform's group/community feature, where communities can have millions of members and their own internal feed, rules, and moderation.

Q47. Design a system for detecting and responding to a coordinated disinformation campaign spreading across multiple accounts and platforms simultaneously during a fast-moving news event.

Q48. Design a system for a platform to let users export their entire personal data archive on request, in compliance with data portability regulations.

Q49. Design a system for a platform to support 'live shopping' events where a creator sells products directly during a livestream, including real-time inventory updates.

Q50. Design a system to support a platform's content recommendation cold-start problem for brand-new users with no engagement history.

E-Commerce & Marketplace Systems

Applied "design X" prompts on e-commerce & marketplace systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q51. Design an e-commerce checkout system that must prevent overselling a limited-stock item during a flash sale with 50,000 concurrent buyers.

Q52. Design a shopping cart system that must persist across devices and sync in real time when a user adds items on mobile and later opens their laptop.

Q53. Design a product search and filtering system for a marketplace with hundreds of millions of listings, supporting complex multi-attribute filters.

Q54. Design a system for calculating and displaying real-time, personalized product recommendations on an e-commerce homepage.

Q55. Design a payment processing system that must handle transactions across 30 different countries with different currencies, tax rules, and payment methods.

Q56. Design a system to handle order fulfillment routing across multiple warehouses, selecting the optimal warehouse for each order based on inventory, distance, and cost.

Q57. Design a system to detect and prevent fraudulent transactions on an e-commerce platform in real time, without unnecessarily blocking legitimate purchases.

Q58. Design a system for managing flash sales where a limited number of a highly desirable product must be sold fairly to the first qualifying buyers.

Q59. Design a system to support dynamic pricing that adjusts product prices in near real-time based on demand, inventory levels, and competitor pricing.

Q60. Design a system for handling product reviews and ratings at the scale of a major marketplace, including detecting fake reviews.

Q61. Design a system to support a marketplace's 'buy now, pay later' financing option, integrating real-time credit decisioning into the checkout flow.

Q62. Design a system for tracking and displaying real-time shipment status to customers, aggregating data from multiple third-party shipping carriers.

Q63. Design a system to handle returns and refunds processing for an e-commerce platform handling millions of orders per month.

Q64. Design a system for a marketplace to support seller-side analytics dashboards showing real-time sales, traffic, and conversion metrics.

Q65. Design a system to handle a scenario where a popular product's price must be updated across millions of already-rendered product pages within seconds of a price change.

Q66. Design a system for a marketplace to support gift cards, including issuance, balance tracking, and redemption across multiple sales channels.

Q67. Design a system to support a subscription-box e-commerce model, handling recurring billing, customization preferences, and shipment scheduling at scale.

Q68. Design a system for detecting and preventing coupon or promo code abuse, such as a single user creating multiple accounts to repeatedly claim a one-time discount.

Q69. Design a system to support real-time inventory synchronization between an e-commerce platform and physical retail store point-of-sale systems.

Q70. Design a system for a marketplace to support multi-vendor shopping carts, where a single checkout can include items from many different independent sellers.

Q71. Design a system to detect when a product listing's price has been set anomalously low, whether due to a genuine seller error or a pricing attack, before it causes a large-scale ordering event.

Q72. Design a system for an e-commerce platform to support 'save for later' and wishlist features that must sync across devices and remain available even for logged-out users.

Q73. Design a system to support same-day delivery logistics, dynamically routing orders to the nearest fulfillment location with available inventory and delivery capacity.

Q74. Design a system for handling checkout abandonment, automatically re-engaging customers who added items to their cart but didn't complete purchase.

Q75. Design a system to support a marketplace's affiliate/referral program, tracking and attributing sales to specific referrers across a multi-touch customer journey.

Q76. Design a system for an online marketplace to detect and delist counterfeit or prohibited product listings at scale.

Q77. Design a system to support real-time price comparison across multiple sellers offering the same product on a marketplace.

Q78. Design a system for handling Black Friday-scale traffic, where load can spike 50x normal volume for a 48-hour window.

Q79. Design a system to support a marketplace's loyalty points program, where points can be earned from purchases and redeemed across a wide range of reward options.

Q80. Design a system for an e-commerce platform to handle a scenario where a payment succeeds on the payment processor's side but the confirmation callback to your system is lost due to a network failure.

Q81. Design a system to support real-time bidding for a business-to-business marketplace where buyers compete for limited supplier inventory through an auction mechanism.

Q82. Design a system for a marketplace to support cross-border sales, handling customs documentation, import duties, and multi-currency pricing display.

Q83. Design a system for handling a scenario where a supplier's inventory feed, integrated via API, is delayed or fails, potentially causing overselling of out-of-stock items.

Q84. Design a system to support a marketplace's 'try before you buy' feature for apparel, including automated return/exchange logistics.

Q85. Design a system for an e-commerce platform to detect and handle 'card testing' fraud, where stolen card numbers are validated through small test transactions.

Q86. Design a system to support a marketplace's price-match guarantee feature, automatically detecting when a competitor offers a lower price on an equivalent product.

Q87. Design a system for handling a scenario where a marketplace must support both instant digital delivery (like software licenses) and physical shipment within the same checkout and order system.

Q88. Design a system to support a marketplace's customer service chatbot that must access real-time order status, inventory, and account data to answer customer questions.

Q89. Design a system for detecting when a marketplace listing's product images have been stolen or copied from another seller without authorization.

Q90. Design a system to support a marketplace's seller onboarding process, including identity verification, tax documentation, and payout account setup, at the scale of onboarding thousands of new sellers daily.

Q91. Design a system for an e-commerce platform to support real-time A/B testing of checkout flow variations, measuring impact on conversion rate.

Q92. Design a system to handle a scenario where a marketplace must instantly and correctly calculate sales tax across thousands of different tax jurisdictions in the United States alone.

Q93. Design a system for a marketplace to support 'buy online, pick up in store' functionality, coordinating real-time inventory between the online platform and specific physical store locations.

Q94. Design a system to detect and prevent 'triangulation fraud,' where a fraudulent seller lists products they don't own, fulfills orders using another retailer's stolen payment information, and pockets the buyer's payment.

Q95. Design a system for an e-commerce platform to support real-time, personalized email marketing campaigns triggered by specific user behaviors, like browsing a product category without purchasing.

Q96. Design a system to support a marketplace's dispute resolution process between buyers and sellers, including evidence submission, case tracking, and automated resolution for straightforward cases.

Q97. Design a system for handling a scenario where a marketplace needs to instantly recalculate and re-display shipping costs as a customer adds or removes items from their cart, factoring in weight, dimensions, and destination.

Q98. Design a system to support cross-selling and upselling recommendations shown specifically during the checkout process, without adding meaningful latency to the purchase flow.

Q99. Design a system for a marketplace to detect when a specific product category is experiencing an unusual return rate spike, potentially indicating a quality or listing accuracy problem.

Q100. Design a system for an e-commerce platform to support real-time stock alerts, notifying customers the moment an out-of-stock item they're waiting for becomes available again.

Fintech & Payments Systems

Applied "design X" prompts on fintech & payments systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q101. Design a payment processing system that must guarantee a transaction is never processed twice, even if the client retries due to a network timeout.

Q102. Design a digital wallet system that must maintain accurate balances across millions of concurrent users with strict consistency guarantees.

Q103. Design a system for real-time fraud detection on credit card transactions that must make a decision within 100 milliseconds.

Q104. Design a system to process bulk payroll payments for a company with 50,000 employees, ensuring every employee is paid correctly and exactly once.

Q105. Design a peer-to-peer money transfer system that must handle the case where the sender has just enough balance, and two transfers are initiated simultaneously.

Q106. Design a system for detecting money laundering patterns across millions of transactions, where individual transactions look legitimate but the aggregate pattern is suspicious.

Q107. Design a system to handle currency conversion for an international remittance service, ensuring the exchange rate shown to the customer matches the rate actually applied.

Q108. Design a system for a stock trading platform that must process buy and sell orders with strict ordering guarantees and sub-millisecond matching latency.

Q109. Design a system to support recurring subscription billing that must handle failed payment retries without double-charging or losing track of a customer's billing cycle.

Q110. Design a system for a buy-now-pay-later service that must make a real-time credit approval decision at the point of checkout for a partner e-commerce site.

Q111. Design a system to reconcile millions of daily transactions between your payment platform's internal ledger and the actual settlement records received from banking partners.

Q112. Design a system for a lending platform to calculate and apply daily interest accrual across millions of active loans with different terms and rates.

Q113. Design a system to detect and block a card-not-present fraud attack where a botnet is testing thousands of stolen card numbers against your payment gateway simultaneously.

Q114. Design a system for a digital bank to support real-time account balance checks that must remain accurate even during a database failover event.

Q115. Design a system to support split payments, where a single purchase amount is divided and charged across multiple different payment methods (like partial gift card, partial credit card).

Q116. Design a system for a cryptocurrency exchange that must maintain accurate, auditable balances while processing thousands of trades per second.

Q117. Design a system to detect a compromised merchant account being used to process fraudulent transactions at unusually high volume.

Q118. Design a system for a payment platform to support chargeback and dispute processing, integrating with card network rules and merchant response workflows.

Q119. Design a system to support real-time currency exchange rate updates for a multi-currency trading platform, ensuring rates are never more than a few seconds stale.

Q120. Design a system for a lending platform to calculate real-time loan eligibility and interest rate offers based on a rapidly changing risk model.

Q121. Design a system to handle a scenario where a payment gateway you depend on experiences a partial outage, affecting only some transaction types.

Q122. Design a system for a financial platform to support real-time net worth or portfolio value calculation across a user's holdings spanning multiple asset classes and accounts.

Q123. Design a system to detect synthetic identity fraud, where fraudsters combine real and fabricated personal information to create a new, fraudulent identity for a financial account application.

Q124. Design a system for a payment platform to support instant, real-time settlement to merchants, as opposed to the traditional multi-day settlement delay.

Q125. Design a system to handle regulatory reporting requirements for a financial platform that must report suspicious transactions to regulators within a strict legal timeframe.

Q126. Design a system for a neobank to support instant account opening with real-time identity verification and initial deposit processing.

Q127. Design a system to support real-time spending limit enforcement for a corporate expense card program, where limits vary by employee, category, and time period.

Q128. Design a system for a financial platform to detect when a customer's transaction pattern suggests they may be a victim of an authorized push payment scam, where they're being manipulated into authorizing a fraudulent transfer themselves.

Q129. Design a system to support micro-lending, processing very small loan amounts to a very large number of borrowers with automated underwriting decisions.

Q130. Design a system for a payment platform to detect and handle 'friendly fraud,' where a legitimate cardholder disputes a charge for an item they actually did receive and knowingly ordered.

Q131. Design a system to support real-time currency hedging recommendations for a business with significant international payment exposure.

Q132. Design a system for a payments platform to support tokenization of card numbers, replacing sensitive card data with non-sensitive tokens throughout the transaction processing pipeline.

Q133. Design a system to support real-time exchange rate arbitrage detection for a multi-currency trading desk, identifying profitable rate discrepancies across different markets before they close.

Q134. Design a system for a financial platform to support real-time credit line utilization tracking across a customer's multiple credit products, ensuring a customer can't exceed their aggregate approved credit through simultaneous transactions across products.

Q135. Design a system to detect a specific type of fraud where an attacker exploits a payment processing race condition to receive goods or services without the corresponding charge actually completing.

Q136. Design a system for a payments platform to support real-time, cross-border remittance with transparent fee calculation shown to the sender before they confirm the transfer.

Q137. Design a system to support automated tax document generation (like 1099 forms) for a payments platform servicing millions of independent contractors and gig workers.

Q138. Design a system for a financial platform to support real-time detection of account takeover attempts, where an attacker has gained unauthorized access to a customer's financial account.

Q139. Design a system to support a robo-advisor platform's automated portfolio rebalancing, executing trades across millions of customer accounts based on their individualized target allocations.

Q140. Design a system for a payments platform to support real-time detection of a merchant potentially engaging in transaction laundering, processing payments on behalf of an undisclosed, higher-risk business.

Q141. Design a system to support real-time, automated expense categorization for a personal finance app, classifying transactions into categories like groceries, dining, and transportation as they occur.

Q142. Design a system for a payments platform to support strong customer authentication requirements (like mandated two-factor authentication for certain transaction types) across the European market, while minimizing friction for lower-risk transactions.

Q143. Design a system to detect when a batch of seemingly unrelated small transactions across many different merchant accounts is actually a coordinated fraud ring testing stolen payment credentials.

Q144. Design a system for a financial platform to support instant peer-to-peer payment requests, where one user can request money from another with real-time notification and one-tap payment.

Q145. Design a system to support a payments platform's real-time currency conversion fee transparency requirement, showing customers exactly how much markup is applied compared to the market mid-rate.

Q146. Design a system for a lending platform to support real-time loan application status tracking, integrating data from multiple internal systems (credit check, document verification, underwriting) into a single customer-facing status view.

Q147. Design a system to support real-time detection of unusual foreign transaction patterns that might indicate a customer's card was compromised while traveling internationally.

Q148. Design a system for a payments platform to support real-time settlement risk management, ensuring the platform doesn't extend more settlement credit to any single merchant than its risk tolerance allows.

Q149. Design a system for a financial platform to support real-time detection of first-party fraud, where a customer opens an account with genuine intent to eventually default or abuse promotional offers rather than committing identity theft.

Q150. Design a system to support real-time multi-party payment splitting for a group expense app, where one payment is divided among several people who each pay their own share independently over time.

Messaging, Communication & Real-Time Systems

Applied "design X" prompts on messaging, communication & real-time systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q151. Design a real-time chat application that must support 500 million concurrent users sending messages with sub-second delivery latency.

Q152. Design a video conferencing system that must support meetings with up to 1,000 participants with acceptable audio and video quality.

Q153. Design a system to guarantee message delivery order in a group chat, even when messages arrive at the server from different senders with network delays.

Q154. Design a push notification system that must reliably deliver notifications to a billion mobile devices across iOS and Android, respecting each platform's specific delivery constraints.

Q155. Design a system for a video calling app to gracefully handle a participant's network connection degrading mid-call without dropping them entirely.

Q156. Design a system to support end-to-end encrypted messaging, where even the messaging platform itself cannot read message content.

Q157. Design a system for a live-streaming platform to minimize latency between a broadcaster's action and it appearing on a viewer's screen, for interactive use cases like live auctions.

Q158. Design a system to support 'typing indicator' and 'online/offline presence' features across a messaging platform with hundreds of millions of users.

Q159. Design a system for a customer support platform to route incoming chat conversations to the most appropriate available agent in real time.

Q160. Design a system to support voice calling over the internet (VoIP) that must maintain call quality across varying network conditions globally.

Q161. Design a system for a messaging platform to support message editing and deletion, ensuring all recipients see the updated or removed content consistently, even if they're offline when the edit happens.

Q162. Design a system to support real-time language translation during a live video call, with minimal added latency.

Q163. Design a system for a messaging platform to detect and prevent spam and phishing links being sent through direct messages at scale.

Q164. Design a system to support screen sharing during a video call for up to 100 participants without significant quality degradation.

Q165. Design a system for an SMS gateway that must reliably deliver text messages globally, routing through different telecom carriers depending on the recipient's country and carrier.

Q166. Design a system to support live polling and Q&A during a large virtual event with tens of thousands of attendees, showing real-time aggregated results.

Q167. Design a system for a messaging platform to support message search across a user's entire message history, spanning years and potentially millions of messages.

Q168. Design a system to detect and handle a scenario where a video conferencing service experiences a sudden surge in meeting creation due to millions of users starting their workday simultaneously in the same time zone.

Q169. Design a system for a messaging platform to support disappearing messages that automatically delete after a set time, ensuring the deletion is enforced consistently across all devices and any local caches.

Q170. Design a system to support real-time collaborative annotation during a video call, where participants can draw or highlight on a shared screen simultaneously.

Q171. Design a system for a large-scale webinar platform that must support one-to-many broadcast to 50,000 attendees with minimal presenter-to-viewer latency.

Q172. Design a system to support real-time voice transcription during a phone call for accessibility purposes, generating live captions with minimal delay.

Q173. Design a system for a messaging platform to handle the scenario where a user has the same account logged into multiple devices simultaneously, ensuring message state (read/unread, delivered) stays consistent across all of them.

Q174. Design a system to support a scalable customer service voice IVR (interactive voice response) system that can handle millions of concurrent calls with dynamic, personalized call routing.

Q175. Design a system for a messaging platform to support group video calls that must gracefully handle participants joining and leaving mid-call without disrupting the experience for others.

Q176. Design a system to support delivery confirmation ('read receipts' and 'delivered' status) for a messaging platform, balancing accuracy against unnecessary network chatter.

Q177. Design a system for a live event platform to support real-time translation of live chat messages into each viewer's preferred language during a global broadcast.

Q178. Design a system to support a walkie-talkie style push-to-talk voice communication feature for a team collaboration app, prioritizing extremely low latency over call quality.

Q179. Design a system for a messaging platform to detect when a conversation has escalated into content that violates platform policy (harassment, threats) and needs urgent moderation attention.

Q180. Design a system to support a scalable, low-latency multiplayer voice chat feature for an online video game, where players in the same virtual space should hear each other with minimal delay.

Q181. Design a system for a video conferencing platform to support recording meetings and generating searchable transcripts afterward.

Q182. Design a system to support real-time collaborative code editing during a technical interview or pair-programming session, including synchronized cursor position and syntax highlighting.

Q183. Design a system for a messaging platform to support broadcast channels where a single sender can send messages to millions of subscribers, but subscribers can't reply directly to the channel.

Q184. Design a system to support real-time noise cancellation and audio enhancement during a video call, processing audio with minimal added latency.

Q185. Design a system for a large-scale conference call platform to support dial-in via traditional phone lines alongside internet-based participants, unifying both into the same call.

Q186. Design a system to support a 'watch party' feature where multiple users can synchronously watch the same video content together with synchronized playback and live chat.

Q187. Design a system for a messaging platform to support message reactions (like emoji reactions) that must update in real time for all conversation participants without excessive server load.

Q188. Design a system to support automatic language detection and routing for a global customer support chat platform, connecting customers to agents who speak their language.

Q189. Design a system for a video conferencing platform to support virtual backgrounds and background blur that work reliably across a wide range of devices with varying processing power.

Q190. Design a system to support a scalable, real-time 'raise hand' and speaker queue feature for a large virtual town hall meeting with thousands of attendees.

Q191. Design a system for a messaging platform to support rich link previews, automatically generating a preview card (title, image, description) when a user shares a URL.

Q192. Design a system to support real-time audio/video call quality monitoring, automatically detecting and alerting on degraded call quality across a platform serving millions of daily calls.

Q193. Design a system for a messaging platform to support forwarding a message to multiple conversations simultaneously, potentially fanning out to thousands of recipients if forwarded to a large group.

Q194. Design a system to support a live audio-only social app (like a live radio/podcast style feature) with millions of listeners tuning into a small number of concurrent broadcasts.

Q195. Design a system for a messaging platform to support scheduled messages that a user composes now but should be delivered automatically at a specified future time.

Q196. Design a system to detect and prevent a distributed denial-of-service style attack specifically targeting a video conferencing platform's signaling servers during a high-profile public event.

Q197. Design a system for a messaging platform to support 'vanish mode' where an entire conversation temporarily switches to non-persistent, auto-deleting messages for the duration of an active chat session.

Q198. Design a system to support live sports commentary overlaid on a video stream, synchronized precisely with the actual game action for millions of concurrent viewers.

Q199. Design a system for a video calling platform to support a 'noise-cancelling background music' feature that lets a presenter play music that's clearly audible to listeners without their own voice being suppressed by the same noise cancellation that would normally treat music as unwanted background noise.

Q200. Design a system for a messaging platform to support a 'do not disturb' scheduling feature that suppresses notifications during specified hours while still delivering messages for later viewing.

Ride-Sharing, Delivery & Location-Based Systems

Applied "design X" prompts on ride-sharing, delivery & location-based systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q201. Design a ride-sharing system that must match riders with the nearest available driver within seconds, across a city with hundreds of thousands of active drivers.

Q202. Design a food delivery system that must estimate accurate delivery times, accounting for restaurant prep time, courier availability, and traffic conditions.

Q203. Design a system for real-time tracking of a delivery driver's location, displayed to the customer on a live map with minimal lag.

Q204. Design a dynamic pricing (surge pricing) system for a ride-sharing platform that adjusts prices in real time based on supply and demand imbalances in a specific geographic area.

Q205. Design a system to optimize delivery route planning for a fleet of hundreds of couriers each carrying multiple orders simultaneously.

Q206. Design a system for a ride-sharing platform to handle a scenario where a rider requests a ride in a low-density rural area with very few available drivers nearby.

Q207. Design a system to detect and prevent GPS spoofing, where a driver or courier fakes their location to appear closer to pickup points than they actually are.

Q208. Design a system for a delivery platform to support real-time inventory checking with thousands of partner restaurants, ensuring customers aren't shown items that are actually out of stock.

Q209. Design a system to calculate optimal driver positioning proactively, suggesting where idle drivers should reposition to maximize their chance of receiving a ride request.

Q210. Design a system for a ride-sharing platform to support ride-pooling, matching multiple riders with overlapping routes into a single shared vehicle trip.

Q211. Design a system to handle driver and rider safety features, like an in-app emergency button that alerts authorities and shares real-time location during an active ride.

Q212. Design a system for a food delivery platform to batch multiple orders from the same restaurant going to nearby delivery addresses into a single courier trip.

Q213. Design a system to support real-time fare estimation shown to a rider before they request a ride, given that actual traffic and route can change during the trip.

Q214. Design a system for a ride-sharing platform to detect and prevent driver-rider collusion fraud, such as fake rides created to earn incentive bonuses.

Q215. Design a system to handle a scenario where a major city event (like a large concert or sports game ending) causes tens of thousands of simultaneous ride requests within a few city blocks.

Q216. Design a system for a delivery platform to support real-time courier earnings calculation, factoring in base pay, distance, time, and any active incentive promotions.

Q217. Design a system to support real-time traffic-aware routing for a fleet of delivery vehicles, dynamically rerouting around unexpected congestion or road closures.

Q218. Design a system for a ride-sharing platform to support scheduled rides booked hours or days in advance, ensuring driver availability is reserved appropriately.

Q219. Design a system to detect anomalous courier behavior suggesting a delivery is at risk of being significantly delayed or not completed, triggering proactive customer communication.

Q220. Design a system for a ride-sharing platform to support multi-stop trips, where a single ride includes several intermediate stops before the final destination.

Q221. Design a system to support a bike-sharing or scooter-sharing platform's real-time vehicle availability map, showing users the location of nearby available vehicles.

Q222. Design a system for a delivery platform to support a 'leave at door' contactless delivery option, including photo confirmation of the delivered package.

Q223. Design a system to detect when a ride-sharing driver's vehicle doesn't match the registered vehicle information on file, a safety-relevant discrepancy.

Q224. Design a system for a food delivery platform to handle a scenario where a restaurant is significantly behind on order prep time during an unexpectedly busy period, risking a cascade of late deliveries.

Q225. Design a system to support real-time carpooling matching for daily commuters with regular, predictable routes, distinct from an on-demand ride-sharing model.

Q226. Design a system for a ride-sharing platform to support accessible vehicle requests for riders with mobility needs, ensuring appropriate vehicle-driver matching.

Q227. Design a system to support real-time delivery zone capacity management, temporarily pausing new order acceptance in a specific area when courier capacity is fully saturated.

Q228. Design a system for a ride-sharing platform to detect when a driver is exhibiting signs of fatigue or unsafe driving behavior based on trip data patterns.

Q229. Design a system to support cross-platform delivery logistics, where a single delivery company fulfills orders originating from multiple different e-commerce platforms simultaneously.

Q230. Design a system for a ride-sharing platform to support real-time fare splitting among multiple riders sharing the same trip.

Q231. Design a system to detect and respond to a scenario where a significant portion of a city's delivery couriers simultaneously go offline during a severe weather event.

Q232. Design a system for a delivery platform to support 'group ordering' where multiple people at the same location can each add items to a single shared order with individual payment.

Q233. Design a system to support real-time estimated wait time display for a ride-sharing platform during a driver shortage, balancing honesty about long waits against not discouraging riders from using the service at all.

Q234. Design a system for a ride-sharing platform to support driver background check verification, ensuring only approved drivers can accept ride requests, with support for periodic re-verification.

Q235. Design a system to support a delivery platform's 'no-contact delivery' compliance during a public health crisis, ensuring couriers and customers can coordinate safely without direct contact.

Q236. Design a system for a ride-sharing platform to detect when a rider's requested destination is likely incorrect or incomplete, before dispatching a driver to a potentially wrong location.

Q237. Design a system to support dynamic delivery fee calculation based on distance, demand, and courier availability, shown transparently to the customer before order confirmation.

Q238. Design a system for a ride-sharing platform to support a rider rating a driver and vice versa after a trip, aggregating ratings while resisting manipulation.

Q239. Design a system to detect and handle 'phantom' ride requests, where a rider requests a ride and cancels repeatedly, potentially to manipulate driver positioning or waste driver time.

Q240. Design a system for a delivery platform to support real-time courier capacity forecasting, predicting how many couriers will be needed in each zone for the next several hours to guide incentive and staffing decisions.

Q241. Design a system to support real-time delivery route optimization that accounts for a courier's mode of transportation (bike, car, on-foot), each with different speed and routing constraints.

Q242. Design a system for a ride-sharing platform to support real-time driver document expiration monitoring, ensuring drivers with expired licenses or insurance are automatically prevented from accepting rides.

Q243. Design a system to support a food delivery platform's real-time 'busy restaurant' warning shown to customers, indicating a restaurant is experiencing unusually high order volume and longer than normal prep times.

Q244. Design a system for a ride-sharing platform to support real-time toll and surcharge calculation for trips that cross toll roads or bridges, ensuring accurate fare billing.

Q245. Design a system to support a delivery platform's real-time 'estimated restaurant rating impact' feature, warning a restaurant partner when consistently slow prep times are likely to hurt their customer ratings.

Q246. Design a system for a ride-sharing platform to support corporate accounts, where a company can manage and pay for its employees' business rides centrally with policy controls.

Q247. Design a system to detect when a courier is at risk of missing a delivery time promise due to an unusually long queue or wait time at the restaurant pickup location.

Q248. Design a system for a scooter-sharing platform to detect vehicles that have been left in a hazardous or improperly parked location, requiring operational team intervention.

Q249. Design a system to support real-time multi-modal trip planning, combining ride-sharing, public transit, and walking directions into a single optimized trip recommendation.

Q250. Design a system for a delivery platform to support automatic re-routing of an order to a different, available courier if the originally assigned courier becomes unresponsive or their device goes offline mid-delivery.

Streaming Media & Entertainment Systems

Applied "design X" prompts on streaming media & entertainment systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q251. Design a video streaming platform that must serve 4K video to 200 million subscribers globally with minimal buffering.

Q252. Design a system to power a video streaming platform's personalized homepage, showing different content rows and rankings for each of hundreds of millions of users.

Q253. Design a system for live sports broadcasting that must deliver a live stream to millions of concurrent viewers with less than 5 seconds of latency from the actual live action.

Q254. Design a system to support video-on-demand content encoding, transcoding a single uploaded video into dozens of resolution and format variants efficiently.

Q255. Design a system for a music streaming platform to generate personalized playlists (like a 'discover weekly' feature) for hundreds of millions of users on a weekly cadence.

Q256. Design a system to detect and block unauthorized redistribution of a live-streamed pay-per-view event to prevent piracy.

Q257. Design a system for a video platform to support seamless quality switching mid-playback as a viewer's network conditions change, without a visible pause or restart.

Q258. Design a system to support closed captioning and subtitles for a video streaming platform, covering dozens of languages generated both by humans and automated speech recognition.

Q259. Design a system for a music streaming platform to accurately calculate and attribute royalty payments to artists based on billions of individual song plays per month.

Q260. Design a system to support a video platform's content recommendation cold-start problem, providing reasonable recommendations for a brand-new user with no viewing history.

Q261. Design a system for a live-streaming platform to handle a scenario where a hugely popular creator's stream suddenly goes viral, causing viewer count to grow from thousands to millions within minutes.

Q262. Design a system to support offline video downloads for a streaming platform's mobile app, allowing users to watch content without an internet connection.

Q263. Design a system for a video platform to detect and automatically remove pirated content uploaded by users, matching against a database of copyrighted reference content.

Q264. Design a system to support a podcast platform's automatic transcription and chapter-marking feature, processing millions of hours of new audio content monthly.

Q265. Design a system for a video streaming platform to support simultaneous multi-device streaming limits, ensuring a subscription's terms (like maximum 2 concurrent streams) are enforced accurately in real time.

Q266. Design a system to support real-time audience measurement and ratings for a live-streamed event, providing advertisers accurate concurrent viewership data during the broadcast.

Q267. Design a system for a music streaming platform to support crossfade and gapless playback transitions between consecutive tracks in a playlist.

Q268. Design a system to support a streaming platform's content licensing window management, automatically making content unavailable in specific regions once a licensing agreement expires.

Q269. Design a system for a video platform to support real-time content moderation of live streams, detecting policy-violating content (violence, prohibited speech) as it's being broadcast.

Q270. Design a system to support a video streaming platform's A/B testing infrastructure for evaluating different video thumbnail images' impact on click-through rate at scale.

Q271. Design a system for a live-streaming platform to support real-time stream health monitoring, automatically alerting broadcasters to issues like dropped frames or audio desync during their live broadcast.

Q272. Design a system to support a music streaming platform's 'listen together' feature, letting friends in different locations listen to the same music synchronized in real time.

Q273. Design a system for a video streaming platform to support dynamic ad insertion into live and on-demand content, personalizing which ads different viewers see within the same stream.

Q274. Design a system to detect when a live stream's video and audio have fallen out of sync during a broadcast, a quality issue that's highly noticeable and disruptive to viewers.

Q275. Design a system for a streaming platform to support binge-watching optimization, automatically preloading the next episode in a series as a viewer approaches the end of their current episode.

Q276. Design a system to support a podcast platform's dynamic ad insertion that must work correctly even for downloaded episodes played entirely offline.

Q277. Design a system for a video platform to support multi-audio-track content, allowing viewers to switch between different language dubs or director's commentary tracks during playback.

Q278. Design a system to support a live-streaming platform's instant replay feature, letting viewers rewind to rewatch a recent moment during an ongoing live broadcast.

Q279. Design a system for a music streaming platform to support real-time collaborative playlists that multiple users can edit simultaneously from different devices.

Q280. Design a system to detect and handle a scenario where a streaming platform's content delivery network experiences a regional outage during a highly anticipated live event.

Q281. Design a system for a video platform to support automatic content categorization and tagging, using AI to analyze uploaded video content and generate relevant genre, mood, and topic tags.

Q282. Design a system to support a streaming platform's 'skip intro' feature, automatically detecting and allowing viewers to skip a TV show's recurring opening title sequence.

Q283. Design a system for a live-streaming platform to support automatic highlight generation, identifying and clipping the most exciting moments from a long live broadcast for later sharing.

Q284. Design a system to support a video platform's regional content licensing compliance, ensuring a piece of content is only playable in countries where the platform holds valid distribution rights.

Q285. Design a system for a music streaming platform to support real-time lyrics display synchronized to the currently playing song's exact timing.

Q286. Design a system to detect when a streaming platform's video encoding pipeline has introduced a quality defect (like visual artifacts) in a specific piece of content before it goes live to viewers.

Q287. Design a system for a video platform to support a 'continue watching' feature that accurately tracks and resumes playback position across all of a user's devices.

Q288. Design a system to support a live-streaming platform's donation/tipping feature during a broadcast, processing real-time payments and displaying on-screen alerts to the streamer with minimal delay.

Q289. Design a system for a streaming platform to support automatic content recommendation explanations, showing users why a specific piece of content was recommended to them (like 'because you watched X').

Q290. Design a system to support a music streaming platform's real-time concert or festival livestream feature, combining live video with synchronized real-time chat and virtual merchandise sales.

Q291. Design a system for a video platform to support region-specific content catalogs, showing different available titles to users depending on their detected location and licensing rights.

Q292. Design a system to support a streaming platform's automatic subtitle translation feature, generating subtitles in a viewer's preferred language for content originally produced in a different language.

Q293. Design a system for a live-streaming platform to support 'multi-view' functionality, letting viewers watch multiple camera angles of the same live event simultaneously.

Q294. Design a system to support a video platform's binge-release strategy, making an entire season of content available simultaneously while managing the storage and CDN pre-warming needed for the anticipated launch-day traffic spike.

Q295. Design a system for a podcast platform to detect and prevent download/stream count fraud, where bad actors artificially inflate a podcast's popularity metrics.

Q296. Design a system to support a streaming platform's simulcast feature, broadcasting the exact same live content simultaneously across multiple different platforms or channels.

Q297. Design a system for a video platform to support accessibility features like audio description tracks for visually impaired viewers, synchronized to the main content.

Q298. Design a system to support a live-streaming platform's viewer poll feature, letting the broadcaster ask their live audience a question and see real-time aggregated results.

Q299. Design a system for a music streaming platform to support smart shuffle, which reorders a playlist for playback in a way that feels random to the listener while avoiding uncomfortable back-to-back repeats of similar-sounding tracks.

Q300. Design a system to support a video platform's watch-party co-viewing feature specifically for mobile devices with potentially unstable cellular connections, ensuring synchronized playback despite variable network conditions between participants.

Search, Recommendation & Discovery Systems

Applied "design X" prompts on search, recommendation & discovery systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q301. Design a web-scale search engine that must index billions of pages and return relevant results within 200 milliseconds.

Q302. Design an autocomplete/typeahead search suggestion system that must return relevant suggestions within milliseconds as a user types each character.

Q303. Design a product recommendation system for an e-commerce platform that must serve personalized recommendations to hundreds of millions of users with sub-50ms latency.

Q304. Design a system to support fuzzy search that can return relevant results even when the user's query contains typos or misspellings.

Q305. Design a system for a job search platform to match job seekers with relevant job postings based on their profile, skills, and application history, at the scale of millions of active postings.

Q306. Design a search ranking system that must balance textual relevance, freshness, and personalization for a news search platform.

Q307. Design a system to support faceted search on an e-commerce platform, where users can filter results by multiple attributes (price, brand, size, color) simultaneously across millions of products.

Q308. Design a system for a video platform to power search across billions of videos using both title/description text and actual video content (like objects or speech within the video).

Q309. Design a system to support 'people also bought' or 'frequently bought together' recommendations at e-commerce scale, updated continuously as new purchase data arrives.

Q310. Design a system for a recipe search platform to support search by available ingredients, returning recipes a user can make with what they currently have on hand.

Q311. Design a system to detect and demote low-quality or spam content from appearing highly in search results, without relying solely on manual review.

Q312. Design a system for a travel booking platform to power search across flights, combining results from hundreds of airlines with real-time pricing and availability.

Q313. Design a system to support voice search, converting spoken queries into accurate search results with the added complexity of speech recognition errors.

Q314. Design a system for a real estate platform to support map-based property search, showing relevant listings as a user pans and zooms across a map interface.

Q315. Design a system to detect and correct for popularity bias in a recommendation system, where already-popular items keep getting recommended more, crowding out lesser-known but potentially relevant items.

Q316. Design a system for a code search platform (like searching across a company's entire codebase) that must return relevant code snippets and definitions within a second, across billions of lines of code.

Q317. Design a system to support 'similar items' recommendations on an e-commerce platform, finding visually and semantically similar products to one a user is currently viewing.

Q318. Design a system for a academic paper search platform to rank results by a combination of relevance, citation count, and publication recency across tens of millions of papers.

Q319. Design a system to support real-time search index updates, ensuring newly published content becomes searchable within seconds rather than waiting for a periodic reindexing batch job.

Q320. Design a system for a music streaming platform to power search that must handle queries for artist names, song titles, and lyrics simultaneously, disambiguating user intent.

Q321. Design a system to detect when a search engine's results for a specific query have degraded in quality, perhaps due to a spam attack or a ranking algorithm regression.

Q322. Design a system for a dating app to power a matching and discovery algorithm that surfaces potential matches based on stated preferences, behavior, and mutual interest signals.

Q323. Design a system to support semantic search that understands the meaning and intent behind a natural language query, not just exact keyword matching.

Q324. Design a system for an image search platform to let users search using an image as the query, finding visually similar images across billions of indexed images.

Q325. Design a system to detect and prevent search result manipulation, where bad actors artificially boost their content's search ranking through coordinated fake engagement.

Q326. Design a system for a marketplace search platform to personalize search result ranking differently for a first-time visitor versus a long-time, highly engaged user.

Q327. Design a system to support search-as-you-browse recommendations, where a user's browsing behavior on a content platform continuously refines what's shown to them within the same session, without requiring an explicit search query.

Q328. Design a system for a legal research platform to support search across millions of case law documents, where exact legal terminology matching is critical alongside broader conceptual relevance.

Q329. Design a system to detect when a recommendation system has developed a harmful filter bubble, repeatedly showing a user increasingly narrow and potentially extreme content based on early engagement signals.

Q330. Design a system for a travel platform to support 'flexible date' search, finding the cheapest flight options across a range of possible departure and return dates rather than a single fixed date.

Q331. Design a system to support search result snippet generation, automatically extracting and highlighting the most relevant portion of a long document to show in search results.

Q332. Design a system for a social platform to power hashtag and topic search, returning the most relevant and recent posts for a given hashtag among potentially millions of posts using it.

Q333. Design a system to detect when a search engine's autocomplete suggestions have started surfacing offensive or inappropriate content due to organic query pattern shifts.

Q334. Design a system for a B2B software marketplace to power search and discovery of software tools, incorporating user reviews, feature comparisons, and integration compatibility into ranking.

Q335. Design a system to support real-time trending search query detection, identifying queries that are suddenly spiking in volume across a search platform.

Q336. Design a system for an online learning platform to recommend the next course a learner should take based on their completed courses, stated goals, and skill gaps.

Q337. Design a system to detect and handle query ambiguity, where a search query could reasonably refer to multiple distinct things (like a search term that's both a movie title and a common phrase).

Q338. Design a system for a recipe platform to support dietary restriction filtering (vegan, gluten-free, allergen-free) that must be highly reliable given the health and safety stakes of an incorrect filter result.

Q339. Design a system to support personalized search ranking for an enterprise internal search tool, where relevance depends heavily on the specific searcher's role, team, and document access permissions.

Q340. Design a system for a video platform to power 'more like this' recommendations shown immediately after a video ends, optimizing for continued watch time within the same session.

Q341. Design a system for a grocery delivery app to power ingredient substitution search, suggesting alternative products when a searched-for item is out of stock.

Q342. Design a system to support cross-language search, letting a user search in one language and retrieve relevant results originally written in a different language.

Q343. Design a system for an app store to rank search results for app discovery, balancing relevance, app quality signals, and preventing pay-to-play manipulation from paid placements overwhelming organic relevance.

Q344. Design a system to support a research platform's citation recommendation feature, suggesting relevant papers a researcher should cite based on the paper draft they're currently writing.

Q345. Design a system for a hotel booking platform to rank search results balancing price, location relevance, user reviews, and the platform's own commercial interests, transparently to the user.

Q346. Design a system to detect when a search platform's spell-correction feature is 'correcting' a query in a way that actually changes the user's clear intent rather than fixing a genuine typo.

Q347. Design a system for a used car marketplace to power search ranking that must balance listing recency, seller trustworthiness, and price competitiveness across millions of listings.

Q348. Design a system to support a podcast platform's search and discovery, allowing users to search not just podcast titles but specific spoken content within episodes.

Q349. Design a system for a fashion e-commerce platform to support visual search, letting users upload a photo of an item they like to find similar products in the catalog.

Q350. Design a system to support a scholarly search platform's 'trending in your field' feature, surfacing papers gaining unusual attention within a researcher's specific area of study recently.

Infrastructure & Developer Platform Systems

Applied "design X" prompts on infrastructure & developer platform systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q351. Design a distributed rate limiter that must enforce consistent limits across thousands of API gateway instances handling millions of requests per second.

Q352. Design a distributed job scheduling system that must reliably execute millions of scheduled tasks across a cluster of worker machines, without duplicate execution.

Q353. Design a CI/CD pipeline system that must build, test, and deploy thousands of independent microservices with minimal cross-team interference.

Q354. Design a distributed logging and monitoring pipeline that must ingest and index log data from tens of thousands of microservice instances in near real time.

Q355. Design a service mesh architecture that must handle service discovery, load balancing, and mutual TLS for thousands of microservices without significant added latency.

Q356. Design a distributed configuration management system that must push configuration updates to thousands of running service instances within seconds of a change.

Q357. Design a container orchestration platform's scheduler that must efficiently pack thousands of container workloads onto available cluster nodes while respecting resource constraints and affinity rules.

Q358. Design a distributed tracing system that must reconstruct the full request path across hundreds of microservices for debugging latency issues.

Q359. Design an internal developer platform that provides self-service infrastructure provisioning for engineering teams without requiring manual operations team intervention.

Q360. Design a secrets management system that must securely distribute and rotate credentials across thousands of services without any service ever seeing another's secrets.

Q361. Design a distributed cache invalidation system that must propagate cache invalidation events to thousands of application server instances within milliseconds of an underlying data change.

Q362. Design a system for automated canary deployment that gradually shifts production traffic to a new service version while automatically rolling back if error rates increase.

Q363. Design a build artifact caching system for a large monorepo, ensuring that only the specific parts of the codebase affected by a change need to be rebuilt.

Q364. Design a distributed lock service that multiple application instances can use to coordinate exclusive access to a shared resource, avoiding both deadlock and split-brain scenarios.

Q365. Design an infrastructure-as-code validation system that must catch dangerous or non-compliant infrastructure changes before they're applied to production.

Q366. Design a distributed rate-limited API gateway that must apply per-customer, per-endpoint rate limits consistently across a globally distributed set of gateway instances.

Q367. Design a system for automated infrastructure cost anomaly detection, alerting engineering teams when their cloud spend deviates significantly from historical patterns.

Q368. Design a health-checking and auto-remediation system that automatically detects and restarts unhealthy service instances across a large fleet without human intervention.

Q369. Design a distributed message queue system that must guarantee at-least-once delivery while supporting millions of messages per second across thousands of topics.

Q370. Design a system for zero-downtime database schema migrations across a fleet of services sharing a large, heavily-used production database.

Q371. Design a system to detect and prevent a 'thundering herd' problem where thousands of services simultaneously retry a failed request to a recovering downstream dependency.

Q372. Design a multi-region deployment orchestration system that must safely roll out application updates across geographically distributed data centers without risking a global outage from a bad deployment.

Q373. Design a system for automated dependency vulnerability scanning across thousands of microservices' codebases, prioritizing which vulnerabilities need urgent remediation.

Q374. Design a distributed rate limiter specifically for protecting a shared, limited third-party API that many internal services need to call.

Q375. Design a system for automated capacity planning, predicting when a service will need additional infrastructure resources based on historical growth trends.

Q376. Design a blue-green deployment system for a stateful service where the database can't simply be duplicated like the stateless application instances.

Q377. Design a system for automated chaos engineering, safely injecting controlled failures into production infrastructure to validate resilience without causing genuine customer impact.

Q378. Design a distributed feature flag system that must let engineers toggle features instantly across millions of running application instances without requiring a redeployment.

Q379. Design a system for automated database backup and point-in-time recovery, capable of restoring a multi-terabyte production database to any point within the last 30 days.

Q380. Design a system to support multi-tenant resource isolation in a shared Kubernetes cluster, preventing one team's workload from starving another's of resources.

Q381. Design a system for automated log retention and archival, balancing debugging value, compliance requirements, and storage cost across petabytes of accumulated log data.

Q382. Design a system for detecting and automatically remediating configuration drift, where a production server's actual configuration has diverged from what infrastructure-as-code declares it should be.

Q383. Design a system for a platform engineering team to provide a standardized, self-service onboarding template for new microservices, ensuring every new service starts with baseline observability, security, and deployment configuration.

Q384. Design a system for automated SSL/TLS certificate management across thousands of internal and external-facing services, ensuring certificates are renewed before expiration without manual tracking.

Q385. Design a distributed system for coordinating a scheduled maintenance window across a fleet of thousands of servers, ensuring patches are applied with minimal service disruption.

Q386. Design a system to detect and prevent a runaway automated process (like a misconfigured script or bot) from making an excessive number of internal API calls that could destabilize shared infrastructure.

Q387. Design a system for automated infrastructure right-sizing, analyzing actual resource utilization across thousands of services to recommend or automatically apply more cost-efficient instance sizing.

Q388. Design a system for distributed rate limiting that must be fair across many API consumers even when the underlying rate-limit state store experiences a brief partial outage.

Q389. Design a system for a platform team to provide standardized incident response tooling, allowing any on-call engineer across the organization to quickly access relevant runbooks, dashboards, and escalation paths during an active incident.

Q390. Design a system to support automated dependency graph analysis across thousands of microservices, helping engineers understand the potential blast radius of a proposed change to a specific service.

Q391. Design a system for automated container image vulnerability scanning integrated directly into the deployment pipeline, blocking deployment of images with critical unpatched vulnerabilities.

Q392. Design a distributed system for coordinating leader election among a cluster of worker nodes, ensuring exactly one leader is active at any time even during network partitions.

Q393. Design a system for automated cost allocation and chargeback reporting, attributing shared cloud infrastructure costs accurately back to the specific teams and services responsible for them.

Q394. Design a system for a platform team to provide a unified API gateway that handles authentication, rate limiting, and request routing for hundreds of internal microservices.

Q395. Design a system for automated infrastructure disaster recovery testing, regularly validating that a full regional failover would actually work without waiting for a genuine disaster to find out.

Q396. Design a distributed system for managing feature flag rollout percentages that must remain consistent for a given user across multiple requests, even as they hit different backend instances.

Q397. Design a system for automated detection of orphaned or unused cloud infrastructure resources that continue to incur cost without providing active value.

Q398. Design a system for a platform team to support canary analysis that automatically determines, using statistical methods, whether a new deployment is safe to fully roll out or should be rolled back.

Q399. Design a system for centralized audit logging that captures every administrative action taken across an organization's entire infrastructure, ensuring the logs themselves can't be tampered with even by someone with elevated infrastructure access.

Q400. Design a system for a platform team to provide standardized observability instrumentation (logging, metrics, tracing) automatically for any service built using the organization's standard service framework, without requiring each team to implement it manually.

Data Platforms & Analytics Systems

Applied "design X" prompts on data platforms & analytics systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q401. Design a real-time analytics dashboard system that must show live metrics (like active users, revenue) updating every few seconds for a platform with billions of daily events.

Q402. Design a data warehouse system that must support both fast, interactive ad-hoc queries from analysts and large, complex batch reporting jobs on the same underlying data.

Q403. Design a system to deduplicate billions of incoming event records from multiple data sources, ensuring accurate downstream aggregate metrics.

Q404. Design an ETL pipeline that must process and transform 10 terabytes of new data daily from dozens of different source systems into a unified data warehouse schema.

Q405. Design a system for a data platform to support real-time change data capture (CDC) from a production transactional database into a downstream analytics system, without impacting production database performance.

Q406. Design a system to support a feature store for machine learning, serving both real-time feature lookups for online inference and batch feature retrieval for offline model training, with consistency between the two.

Q407. Design a system for computing accurate, deduplicated unique visitor counts across a website receiving billions of page views daily.

Q408. Design a data lake architecture that must store both structured and unstructured data from hundreds of different internal systems while remaining genuinely discoverable and query-able, avoiding becoming an ungoverned data swamp.

Q409. Design a system to support real-time fraud detection scoring that requires joining streaming transaction data with historical customer behavior data within milliseconds.

Q410. Design a system for computing complex, multi-dimensional business metrics (like revenue by region, product, and customer segment simultaneously) that must be queryable interactively by business analysts.

Q411. Design a system to detect data quality issues (missing values, out-of-range values, schema violations) automatically across hundreds of data pipelines before bad data reaches downstream consumers.

Q412. Design a system for a data platform to support self-service SQL querying for hundreds of internal analysts, ensuring query performance remains acceptable despite widely varying query complexity and skill levels.

Q413. Design a system to compute real-time A/B test result significance, continuously updating statistical confidence as new experiment data arrives.

Q414. Design a system for a data platform to support reprocessing historical data when a bug is discovered in a transformation pipeline that's been running incorrectly for months.

Q415. Design a system to support real-time anomaly detection across thousands of business metrics simultaneously, alerting when any metric deviates significantly from its expected pattern.

Q416. Design a system for computing accurate session-based analytics (like average session duration, pages per session) from raw clickstream event data at web scale.

Q417. Design a system to support a machine learning training pipeline that must efficiently process petabytes of training data distributed across a cluster of GPU workers.

Q418. Design a system for computing real-time cohort retention analysis, tracking what percentage of users from each signup cohort remain active in subsequent weeks.

Q419. Design a system to support a data platform's data lineage tracking, letting engineers trace exactly which upstream sources and transformations produced any given downstream dataset or metric.

Q420. Design a system for real-time clickstream analytics that must handle a scenario where a significant fraction of events arrive late due to mobile devices being temporarily offline.

Q421. Design a system to support multi-tenant data isolation in a shared data warehouse, ensuring one customer's data and queries can never be accessed by or impact another customer sharing the same underlying infrastructure.

Q422. Design a system for computing real-time inventory forecasting, predicting future stock needs based on historical sales patterns, seasonality, and current trends.

Q423. Design a system to support a data platform's approach to handling personally identifiable information across a complex pipeline, ensuring compliance with data deletion requests even for data that's been copied into multiple derived datasets.

Q424. Design a system for computing real-time supply chain visibility, aggregating data from thousands of suppliers, warehouses, and logistics partners into a unified tracking view.

Q425. Design a system to support real-time detection of data pipeline failures, alerting engineers before a broken pipeline causes stale or missing data to reach business-critical dashboards.

Q426. Design a system for a data platform to compute accurate customer lifetime value predictions, combining historical purchase data with predictive modeling.

Q427. Design a system to support real-time geographic heat-map visualization of user activity across a global platform, updating as new activity events stream in.

Q428. Design a system for a data platform to detect when an upstream data source has silently changed its schema or data format, potentially breaking downstream pipelines without an obvious error.

Q429. Design a system to compute real-time marketing attribution, determining which marketing touchpoints contributed to a customer's eventual conversion across a multi-week, multi-channel customer journey.

Q430. Design a system for a data platform to support efficient incremental backup of a multi-petabyte data warehouse, minimizing both backup time and storage cost.

Q431. Design a system to support real-time supply and demand matching analytics for a marketplace platform, continuously computing the current balance of available supply against active demand across thousands of micro-markets.

Q432. Design a system for computing real-time customer churn risk scores, identifying customers likely to cancel their subscription before they actually do.

Q433. Design a system to support a data platform's approach to computing accurate, auditable financial reporting metrics that must reconcile exactly with the underlying transactional source of truth.

Q434. Design a system for real-time website performance monitoring, aggregating page load time and error rate data from millions of real user sessions globally.

Q435. Design a system to support computing real-time leaderboards for a large-scale online competition or game, ranking millions of participants with frequently updating scores.

Q436. Design a system for a data platform to detect and prevent a scenario where a scheduled batch job's output silently regresses in quality due to an upstream data issue, without anyone noticing until a business stakeholder complains.

Q437. Design a system to support real-time computation of dynamic leaderboard-based incentive payouts for a gig economy platform, where workers earn bonuses based on their relative ranking against peers during a specific time window.

Q438. Design a system for computing real-time price elasticity analysis, determining how sensitive demand is to price changes for thousands of individual products.

Q439. Design a system to support computing real-time aggregate carbon footprint metrics for a logistics company, based on live vehicle telemetry and route data across a large delivery fleet.

Q440. Design a system for a data platform to support computing real-time customer segmentation, dynamically grouping customers into behavioral segments as their activity patterns evolve.

Q441. Design a system to compute real-time supply chain disruption risk scoring, aggregating signals from news feeds, weather data, and supplier performance history to flag at-risk shipments proactively.

Q442. Design a system for a data platform to support computing accurate, GDPR-compliant data processing activity records, documenting exactly what personal data is processed, why, and by which systems.

Q443. Design a system to compute real-time website conversion funnel analytics, showing where users drop off across a multi-step signup or checkout process.

Q444. Design a system for a data platform to support computing real-time inventory shrinkage detection, identifying discrepancies between expected and actual physical inventory levels that might indicate theft or process errors.

Q445. Design a system to compute real-time content engagement scoring across a media platform, ranking which pieces of content are currently generating the most meaningful engagement beyond simple view counts.

Q446. Design a system for a data platform to support computing real-time employee productivity and workload analytics for a large distributed workforce, balancing genuine business insight against privacy and surveillance concerns.

Q447. Design a system to compute real-time predictive maintenance alerts for industrial equipment, using sensor telemetry data to predict failures before they occur.

Q448. Design a system for a data platform to support computing accurate real-time weather-adjusted sales forecasting for a retail chain, accounting for how weather conditions affect foot traffic and purchasing patterns differently across product categories.

Q449. Design a system to compute real-time social sentiment analysis aggregating and scoring public sentiment toward a brand across social media, news, and review platforms simultaneously.

Q450. Design a system for a data platform to support computing real-time cross-device user identity resolution, recognizing that the same individual is using a platform across their phone, laptop, and tablet despite no explicit login linking them.

AI/ML & Emerging Systems

Applied "design X" prompts on ai/ml & emerging systems. In the room, expand each with a rough capacity estimate, a component sketch and the follow-up trade-offs the interviewer pushes on.

Q451. Design a system to serve a large language model for a chat application handling millions of concurrent conversations with low latency.

Q452. Design a system for a retrieval-augmented generation (RAG) pipeline that must find and incorporate relevant documents into an LLM's context from a knowledge base of millions of documents.

Q453. Design a system to detect and prevent prompt injection attacks against an LLM-powered customer service application that has access to sensitive backend systems.

Q454. Design a system for real-time content moderation using a machine learning model that must classify millions of pieces of user-generated content per hour with very low latency.

Q455. Design a machine learning model training pipeline that must retrain a fraud detection model daily using the previous day's confirmed fraud outcomes as new labeled data.

Q456. Design a system to serve real-time personalized recommendations using a machine learning model, ensuring the model's predictions stay fresh as user behavior evolves throughout a session.

Q457. Design a system for detecting AI-generated (deepfake) images and videos being uploaded to a content platform at scale.

Q458. Design a system to support A/B testing of different machine learning model versions in production, safely comparing a new model's performance against the current production model.

Q459. Design a system for an AI coding assistant that must provide code completions with sub-200ms latency while maintaining awareness of the user's entire codebase context.

Q460. Design a system to detect when a production machine learning model's performance has silently degraded due to data drift, where the real-world data distribution has shifted from what the model was trained on.

Q461. Design a system for an AI agent that can autonomously execute multi-step tasks (like booking travel) by calling external APIs, ensuring it doesn't take harmful or costly unintended actions.

Q462. Design a system to support real-time content recommendation for a news platform that must balance personalization with avoiding algorithmic filter bubbles and misinformation amplification.

Q463. Design a system for serving a machine learning model that requires GPU inference, optimizing for cost efficiency given GPU compute's high price relative to CPU.

Q464. Design a system to detect hallucinations in an LLM's output for a use case where factual accuracy is critical, like a medical information chatbot.

Q465. Design a system for a machine learning feature engineering pipeline that must compute hundreds of features from raw event data, keeping both batch training data and real-time serving features consistent.

Q466. Design a system to support real-time voice assistant query understanding, converting spoken natural language into a structured action the system can execute, with minimal latency.

Q467. Design a system for detecting bias in a machine learning model's predictions across different demographic groups, and correcting for it without significantly sacrificing overall accuracy.

Q468. Design a system to support efficient fine-tuning of a large language model for a specific domain, given the significant computational cost of full model retraining.

Q469. Design a system for real-time anomaly detection in network security logs using machine learning, distinguishing genuine security threats from normal but unusual network behavior.

Q470. Design a system to support multi-modal AI model serving, where a single model must process combined text, image, and audio input for tasks like analyzing a video with commentary.

Q471. Design a system for an AI-powered code review assistant that must analyze pull requests and flag potential bugs or security issues without significantly slowing down a team's development velocity.

Q472. Design a system to detect when an AI model deployed for automated decision-making (like loan approval) is making decisions that would violate fair lending regulations, before those decisions are finalized.

Q473. Design a system for serving personalized AI-generated content (like personalized marketing copy) at scale, ensuring generated content meets brand and quality standards before reaching customers.

Q474. Design a system to support efficient vector similarity search across billions of embeddings for a large-scale recommendation or semantic search system.

Q475. Design a system for detecting when a machine learning model's training data may have been poisoned with malicious or intentionally corrupted examples.

Q476. Design a system to support real-time AI-powered translation for a customer service platform, handling both text chat and voice calls across dozens of language pairs.

Q477. Design a system for an AI-powered fraud detection model that must explain its decisions to human reviewers and, where legally required, to affected customers.

Q478. Design a system to support continuous evaluation of a production LLM application's output quality, catching regressions introduced by prompt changes, model updates, or underlying data shifts.

Q479. Design a system for an AI-powered virtual try-on feature for an e-commerce platform, rendering how a clothing item would look on a customer's uploaded photo in near real time.

Q480. Design a system to support real-time AI-powered document processing, extracting structured data (like invoice line items) from a wide variety of unstructured document formats and layouts.

Q481. Design a system for detecting when an AI chatbot is being used for a purpose outside its intended scope (like a customer service bot being manipulated into providing unrelated harmful content).

Q482. Design a system to support efficient batch inference for a machine learning model that needs to score millions of records overnight, minimizing total processing time and cost.

Q483. Design a system for an AI model that must make real-time content recommendations while respecting a rapidly changing set of content licensing restrictions (like content becoming unavailable in specific regions).

Q484. Design a system to support real-time detection of AI-generated spam or fake reviews flooding a platform, distinguishing them from genuine human-written content.

Q485. Design a system for an AI-powered personal assistant that must maintain conversational context across a long-running interaction while managing the practical limits of an LLM's context window.

Q486. Design a system to detect and handle a scenario where an AI content generation tool is being used to mass-produce low-quality, search-engine-manipulation-focused content at scale.

Q487. Design a system for an AI-powered meeting summarization tool that must generate accurate summaries from meeting transcripts while handling the ambiguity and informality of natural spoken conversation.

Q488. Design a system to support cost-efficient serving of multiple different fine-tuned versions of the same base large language model, each customized for a different enterprise customer.

Q489. Design a system for detecting when a computer vision model used for quality control on a manufacturing line has started producing systematically incorrect classifications due to a physical change like a camera going slightly out of focus.

Q490. Design a system for an AI-powered resume screening tool that must efficiently process thousands of applications per job posting while avoiding illegal discriminatory bias against protected characteristics.

Q491. Design a system to support real-time AI-powered live sports commentary generation, automatically generating natural language commentary synchronized with live game action.

Q492. Design a system for detecting when an AI customer service agent has reached the limits of its ability to help and should hand off to a human agent, without frustrating the customer through a poor handoff experience.

Q493. Design a system to support efficient serving of an AI image generation model, handling a highly variable request queue where generation time varies significantly based on requested image complexity and resolution.

Q494. Design a system for an AI-powered legal document review tool that must flag potentially problematic clauses in contracts while working across many different contract types and jurisdictions.

Q495. Design a system to detect when an AI voice cloning or synthesis tool is being used to create unauthorized deepfake audio impersonating a real person.

Q496. Design a system for an AI-powered supply chain optimization tool that must recommend inventory and logistics decisions while accounting for genuine uncertainty in demand forecasts.

Q497. Design a system to support AI-powered real-time video call transcription and action-item extraction, generating a structured summary of decisions and follow-ups immediately after a meeting ends.

Q498. Design a system for detecting when a recommendation algorithm's use of AI-generated content in its candidate pool is disproportionately displacing human-created content, potentially harming a content platform's creator ecosystem.

Q499. Design a system to support an AI-powered educational tutoring platform that must adapt its teaching approach in real time based on a student's demonstrated understanding during a session.

Q500. Design a system for an AI content safety classifier that must operate across dozens of languages and cultural contexts with consistent policy enforcement despite significant linguistic and cultural nuance.

System Design Interview – For Architects

Related News

What Is a Single Point of Failure (SPOF)?

What Is a Health Check?