Study interactive :: Progress tools open in the Study Hub reader.

System Design for Beginners

Notes I'm writing as I go through "System Design for Beginners" (31 lessons: 22 foundations + 9 interview-style designs). Each file is a chapter you can read on its own. No video required.

The goal is not to memorize buzzwords. By the end you should be able to look at a product like Twitter or Uber and have a rough idea of what's running behind the scenes, and why.

Reading style is borrowed from Alex Xu's books: short sentences, real numbers, concrete examples, trade-offs spelled out.

Where this fits in the repo

This folder is a foundational side track that pairs with the main Road to Machine Learning curriculum. It teaches the general backend vocabulary the ML guides assume you know.

You can read this in parallel with the numbered modules. It pairs especially well with module 13 (deployment) and 14 (MLOps).

Diagrams

Topology and flow figures in these lessons use Mermaid fences (```mermaid) instead of ASCII boxes.

On GitHub you see the Mermaid source. On the Study Hub reader they bake to SVG at sync time (Fit / zoom / pan). Hash rings in 13. Consistent Hashing are drawn as a clockwise walk flattened left-to-right. Mermaid is weak at true circles.

The same pattern is used outside this folder for high-ROI pipelines and decision trees (RAG, ML serving, ETL / ER, MLflow, Power BI, MLOps, and several quick-reference pickers). Directory trees (├──) stay ASCII on purpose.

How to read this

Go in order if you're new. The first few chapters set up the vocabulary the later ones lean on.

If you already work in software, jump around. The CAP, sharding, and message queue chapters are the ones engineers usually come back to during interviews.

Lessons

Background

Networking

APIs

Caching

Proxies

Storage

Big Data

Capstone

Interview-style designs

Practice problems using the four-step process from Chapter 21. Each one is original, written from scratch, with the same structure: clarify, estimate, high-level, deep dives.

Books and resources I keep coming back to

A note on numbers

When I quote a number like "Redis handles 100k ops/sec" or "S3 stores 11 nines of durability", those numbers are real-world ballparks. They move year to year. Use them to get a sense of scale, not as gospel.