Contents
Course · Home
How do big systems really work, and how do I explain them in an interview?
A self-study course for SDE-2 system design interviews. Every page teaches with pictures first. Then it shows what actually happens underneath, with official sources. It ends with interview questions and model answers you can hide and reveal.
TL;DR
- Read the pages in order. Each one builds on the one before.
- Every page has the same shape: words you need → concepts with pictures → trade-offs → traps → interview questions → cheat sheet.
- Answer each interview question out loud before you click "Show answer".
- Finish with page
15(how to run the interview) and page16(11 classic designs, worked end to end).
16pages
one topic per page
5parts
foundations → building blocks → distributed → cloud → interview
11designs
classic problems worked end to end on page 16
—The learning path
In everyday words
Like learning to cook. First knives and heat, then single dishes, then a full meal for a hundred people, then cooking live in front of a judge.Pages
1Foundations
Page 01
Scaling up vs out, p99 latency, "nines" of availability, SLOs, and back-of-the-envelope maths.
Page 02
What happens when you type a URL. TCP vs UDP, HTTP/1.1 to HTTP/3, REST, gRPC, GraphQL, WebSockets.
2Building blocks
Page 03
L4 vs L7, balancing algorithms, health checks, token bucket and sliding windows.
Page 04
Cache-aside vs write-through, eviction, stampedes, HTTP cache headers, edge vs origin.
Page 05
Why it is fast, data structures, persistence, Cluster hash slots, locks, Streams.
Page 06
ACID, isolation levels, B-tree vs LSM, NoSQL families, how to pick one.
3Distributed systems
Page 07
Leader-follower, quorums, shard keys, hot partitions, virtual nodes.
Page 08
CAP and PACELC, consistency models, Raft, clocks, 2PC, retries with jitter.
Page 09
Queue vs pub/sub, Kafka partitions and groups, delivery promises, DLQs.
Page 10
Circuit breaker, saga, outbox, CQRS, service discovery, canary deploys.
Cloud and the interview
4Cloud and operations
Page 11
Block vs file vs object storage, pre-signed uploads, inverted indexes, Bloom filters.
Page 12
Regions and AZs, EC2 and Auto Scaling, ALB vs NLB, Lambda, VPC, IAM, S3.
Page 13
RDS vs Aurora, DynamoDB in depth, SQS vs SNS vs EventBridge vs Kinesis.
Page 14
Logs, metrics, traces, SLO alerts, RPO vs RTO, OAuth, JWT, password hashing.
5The interview
Page 15
What SDE-2 interviewers score, a minute-by-minute plan, and how to talk about trade-offs.
Page 16
URL shortener, rate limiter, chat, news feed, notifications, KV store, crawler, typeahead, YouTube, Uber, payments.
Plan
—Study plan
| Week | Read | Practice |
|---|---|---|
| Week 1 | Pages 01-04 | Do the estimation example on page 01 again without looking. Answer every question on pages 01-04 out loud. |
| Week 2 | Pages 05-08 | Explain Redis, isolation levels, consistent hashing and CAP to a friend in 2 minutes each. |
| Week 3 | Pages 09-14 | Draw the "3-tier web app on AWS" from memory. Explain Kafka vs a queue. |
| Week 4 | Pages 15-16 | Run 2 timed mock interviews per week using the framework. Use a new design from page 16 each time. |
—How to use each page
Words you need
Read these first. Every hard word on the page is explained here.
Picture → everyday words → Actually
The picture gives the idea, the comparison makes it stick, and "Actually" gives the real mechanism with its source.
Interview questions
Answer out loud first, then click "Show answer". Note the ones you missed and come back to them.
Cheat sheet
The numbers to remember. Look at it again the day before the interview.
Built 2026-10-25. Facts on each page are checked against official docs, which are listed in that page's Sources section.