System Design

Design systems the way production engineers do: start with goals and constraints, make trade-offs explicit, draw clear diagrams, then prove the critical path with code.

This course is for working software engineers and people who want to become software engineers. You do not need to be a cloud architect first. You need curiosity about how requests, data, and failures move through a system.

Who this is for

You are…	You will practise…
Aspiring software engineer	Turning vague features into APIs and data models
Backend / full-stack engineer	Scaling paths, consistency, and failure modes
DevOps / platform engineer	Seeing services as products with SLOs and blast radius
Interview candidate	Structured design conversations with evidence

What you will learn

• Clarify functional and non-functional requirements before drawing boxes
• Estimate capacity with simple, defensible maths
• Choose architecture styles for the problem — not fashion
• Trace a request from client through DNS, CDN, load balancer, and service
• Explain trade-offs (latency vs consistency vs cost) in plain language
• Implement thin but real paths in Python so designs stay honest

Course structure

Part	Focus	Status
A · Foundations	Design process, quality attributes, styles, request path	Modules 1–4 ready
B · Building blocks	Storage, cache, messaging, APIs, resilience	Modules 5–9 ready
C · Classic systems	Shortener, feed, uploads, search	Modules 10–13 ready
D · Realtime	Chat, notifications, collaboration, capstone	Modules 14–17 ready

Start here

1. How to design a system — the thinking process
2. Quality attributes and trade-offs — what “good” means
3. Application architecture styles — monolith to events
4. Client, edge, and service path — where a request travels
5. Data storage — access patterns and engines
6. Caching — layers, TTL, stampede control
7. Messaging and async — queues and idempotency
8. APIs and communication — contracts and timeouts
9. Observability and resilience — SLOs and failure modes
10. URL shortener — redirect-optimised design
11. News feed / timeline — fan-out trade-offs
12. File / media upload — direct upload pipelines
13. Search / autocomplete — indexes and suggest
14. Realtime chat — WebSockets and durable fan-out
15. Notifications and presence — heartbeats and routing
16. Collaborative and streaming patterns — OT/CRDT intuition
17. Capstone — collaboration board — integrate the course

See the roadmap for the full journey.

How each tutorial works

Every lesson follows the same arc:

Requirements → Trade-offs → Architecture (diagrams) → Python lab → Interview check

Diagrams are first-class. Labs use Python so you can run and break things on your machine.

Prerequisites

• Comfortable with HTTP (methods, status codes, JSON)
• Can write small Python scripts
• Basic SQL ideas (tables, primary keys)
• Git installed

All course labs run with Python alone. Docker/Redis are optional upgrades for portfolio demos.