Case Study Index: Design Classics, Grouped by Theme

Between the module chapters and the five worked examples you just read, this course has now bridged a chapter to a specific level about twenty-five times, and the HLD Lab has 100 levels. Everything beyond the levels already linked is capstone territory: applying the whole toolbox at once, with less hand-holding about which concept applies where. Here's a map of what's left, grouped by theme rather than listed in order.

• -Pick the theme that matches what you are preparing for, not the next number in the list.
• -Skim the chapters named under that theme first, so the toolbox is fresh.
• -Attempt a level with the 5-step loop from the previous chapter: clarify, estimate, sketch, find the bottleneck, name the trade-off.
• -Read the review after your run. It shows what your design missed and why, and that is where most of the learning happens.

The systems most frequently asked about in real interviews, each combining several modules from this course at once: Photo Sharing (Level 16), Payment System (Level 21), Distributed Message Queue (Level 22), Team Messaging (Level 31), and Email Service (Level 40).

Revisit first: object storage and CDNs for photo sharing, idempotency and sagas for payments, delivery guarantees for a message queue, WebSockets and fan-out for messaging, and queues and search for email.

Systems with a genuinely hard core problem, not just more scale: Stock Exchange Order Book (Level 28), Object Storage / S3 Clone (Level 33), Code Repository (Level 35), Maps & Navigation (Level 39), Video Conferencing (Level 45), and Auction System (Level 47).

Revisit first: replication and consistency, sharding, object storage and real-time systems. These problems usually have one hard core, such as ordering, durability or geometry, and the rest is the toolbox you already know.

Levels 55 through 59 deliberately revisit five systems from earlier in the lab (Ride-Sharing, Video Streaming, Real-Time Messaging, Collaborative Editing, and a Distributed Task Scheduler) this time at Expert and Master difficulty, expecting the full toolbox this course has built rather than the beginner's version of the same problem.

Revisit first: the same chapters as your first attempt, plus whatever the review told you that you missed. Compare the two attempts: the difference is your progress.

The Grandmaster and Master tier levels are where every module in this course shows up in one design at once: Global Stock Exchange (Level 60), Multi-Tenant SaaS Platform (Level 61), Global Payment Processing (Level 63), Distributed Tracing Platform (Level 64), Distributed Search Engine at Google scale (Level 67), Multi-Region Active-Active Database (Level 68), and Real-Time Fraud Detection (Level 69).

Revisit first: multi-region and geo-sharding, observability, and the resilience chapters. At this scale the question is less how to build it and more what happens when part of it fails.

Levels 50 through 54 apply this same system-design toolbox to AI infrastructure specifically: a RAG Pipeline, a Vector Search Engine, an AI Gateway, a Real-Time Recommendation Engine, and an Autonomous Agent Platform. If these interest you, ScaleDojo's GenAI lab track goes considerably deeper.

Levels 71 through 100 are a second, later addition to the lab covering systems this course doesn't walk through directly - the everyday backbone of real products (Webhook Delivery, Distributed Saga Orchestrator, Flash Sale Inventory, Subscription Billing) through Levels 71-85, then planet-scale and frontier problems (Distributed Ledgers, Smart Grid Telemetry, Global Supply Chain Tracking) through Levels 86-100, closing with a second, harder capstone at Level 100.