🗺️ System Design Mastery — Complete Guide Life Flow

A visual map of the entire curriculum — understand what to learn, in what order, and why. Whether you're a beginner, a backend engineer, an interview candidate, or a researcher, this page is your starting point.

🌐 Big Picture — What Is This Repository?

┌─────────────────────────────────────────────────────────────────────────┐
│                    SYSTEM DESIGN MASTERY (SDM)                          │
│                                                                         │
│  A free, open-source documentation platform that teaches you how to    │
│  design the systems powering Google, Netflix, Uber, WhatsApp & more.   │
│                                                                         │
│  📚 12 Levels  ·  88+ Markdown Files  ·  18,000+ Lines of Content      │
│  🏗️ 14 Real-World Systems  ·  100+ Mermaid Diagrams  ·  Live on Vercel │
└─────────────────────────────────────────────────────────────────────────┘

🔭 Full Curriculum Overview (All 12 Levels)

🎯 Choose Your Learning Path

🟢 Path A — The Beginner (0–2 years experience)

Goal: Understand how the internet and large-scale systems actually work.
Time: 4–6 weeks · 2–3 hours/week

Key topics to focus on:

networking.md — TCP/IP, HTTP/HTTPS, DNS basics
what-happens-when-you-type-google.md — end-to-end request journey
high-level-architecture.md — boxes, arrows, and components
scaling-1-to-1-billion.md — how systems grow
database-management.md — when to use SQL vs NoSQL

🟡 Path B — The Intermediate Engineer (2–5 years)

Goal: Master distributed systems, event-driven design, and backend complexity.
Time: 3–4 weeks · 4–5 hours/week

Key topics to focus on:

database-sharding.md — shard strategies, hot spots
cdn-optimization.md — edge caching, cache-control headers
dead-letter-queues.md — resilient async processing
grpc.md — binary protocol, protobuf, streaming
circuit-breakers.md — fault isolation in microservices
saga-pattern.md — distributed transactions

🔴 Path C — The Interview Candidate (2–3 weeks sprint)

Goal: Ace the FAANG system design interview.
Time: 2–3 weeks · 5–6 hours/week

Must-practice systems (Level 7):

Priority	System	Key Concept
⭐⭐⭐	URL Shortener	Hashing, redirection, analytics
⭐⭐⭐	Twitter / X	Fan-out, timeline, feed ranking
⭐⭐⭐	WhatsApp	WebSocket, message ordering, E2E
⭐⭐⭐	Uber	Geo-spatial, matching, real-time
⭐⭐⭐	Netflix	CDN, video encoding, recommendation
⭐⭐	Notification System	1B/day, fan-out, multi-channel
⭐⭐	Instagram	Photo storage, CDN, social graph
⭐⭐	YouTube	Upload pipeline, HLS streaming
⭐	Google Search	Crawling, indexing, ranking
⭐	TikTok	Recommendation engine, ML pipeline

🟣 Path D — The Tech Lead / Architect

Goal: Design production-grade, observable, and resilient systems.
Time: Ongoing · 2–3 hours/week

🤖 Path E — The AI / ML Engineer

Goal: Architect AI-powered systems with LLMs, RAG, and multi-agent pipelines.
Time: 2–3 weeks · 3–4 hours/week

📦 What's Inside Each Level

Level 1 — Foundations

The bedrock. Never skip this.

File	What You Learn
`networking.md`	TCP/IP, UDP, HTTP/HTTPS, TLS handshake
`dns-lookup.md`	Full DNS resolution from browser to root
`load-balancing-algorithms.md`	Round Robin, Consistent Hashing, Least Connections
`protocols-communication.md`	REST, WebSockets, gRPC, MQTT — when to use each
`high-level-architecture.md`	Components: LB, DB, Cache, CDN, Queue
`design-patterns.md`	Proxy, Reverse Proxy, Sidecar, Ambassador
`what-happens-when-you-type-google.md`	Full end-to-end request journey
`fault-tolerance-crashes.md`	Graceful degradation, failover
`clustering-fundamentals.md`	Active-passive vs active-active
`native-mobile-stacks.md`	iOS, Android architecture patterns

Level 2 — Scalability

File	What You Learn
`scaling-1-to-1-billion.md`	8 growth stages with trade-offs at each tier

Key concepts: Horizontal scaling, vertical scaling, CAP Theorem, consistency models, stateless services.

Level 3 — Databases & AI Data Systems

File	What You Learn
`database-management.md`	SQL vs NoSQL, when to use each
`database-sharding.md`	Range, hash, directory-based sharding
`vector-databases.md`	Storing & searching meaning with Pinecone/Weaviate
`vector-embeddings.md`	How text → vectors, cosine similarity, chunking
`ann-hnsw-index.md`	How HNSW finds nearest neighbors in milliseconds
`similarity-metrics.md`	Euclidean vs Cosine vs Dot Product
`faiss-vector-database.md`	Meta's billion-scale vector search library
`rag-pattern.md`	Retrieval-Augmented Generation end-to-end
`llm-ai-agents.md`	Transformer internals, AI Agent loops, tool use, multi-agent systems

Level 4 — Caching

File	What You Learn
`cdn-optimization.md`	CDN topology, cache-control, purging strategies

Key concepts: Cache-aside, write-through, write-behind, TTL, Redis data structures, eviction (LRU/LFU).

Level 5 — Messaging & Queues

File	What You Learn
`README.md`	Kafka, RabbitMQ, SQS — when to pick which
`dead-letter-queues.md`	DLQ patterns, retry strategies, poison pill handling

Key concepts: Pub/Sub, consumer groups, at-least-once vs exactly-once delivery, backpressure.

Level 6 — Microservices

File	What You Learn
`grpc.md`	Protobuf, streaming RPCs, load balancing
`role-menu-permission.md`	Enterprise RBAC/ABAC access control design

Key concepts: API Gateway, service mesh, sidecar proxy, service discovery, bulkhead pattern.

Level 7 — Real-World Systems (14 deep dives)

Level 8 — Case Studies & Interview Prep

File	What You Learn
`01-url-shortener.md`	Full interview walkthrough — 8 steps
`02-chat-application.md`	Real-time messaging design interview
`03-netflix-video-streaming.md`	Video platform interview strategy
`how-to-manage-case-studies.md`	How to structure your own case studies

Level 9 — Design Patterns

File	What You Learn
`circuit-breakers.md`	Fault isolation — Open/Half-open/Closed states
`saga-pattern.md`	Distributed transactions without 2PC
`event-sourcing.md`	Append-only event log as source of truth
`idempotency-keys.md`	Safe retries in payment and messaging systems
`exponential-backoff.md`	Smart retry logic with jitter
`system-design-connection.md`	Connecting patterns to real-world decisions

Level 10 — Monitoring & Observability

File	What You Learn
`latency-percentiles.md`	P50 / P95 / P99 — why averages lie
`error-rate-alerts.md`	Setting meaningful alert thresholds
`realtime-dashboards.md`	Grafana-style dashboard design
`transaction-success-rate.md`	SLI/SLO for payment critical paths

Level 11 — Academic Research

File	What You Learn
`README.md`	How to find a research niche, read papers, publish

Seminal Papers: Google Bigtable, Dynamo, MapReduce, Spanner, Raft, Kafka — the papers that built the internet.

Level 12 — Interview Prep (20 deep-dive topic guides)

Located in 12-interview-prep/ — the most comprehensive section for FAANG prep.

Guide	Topic
`requirements-clarification.md`	How to ask the right questions
`back-of-envelope-estimation.md`	DAU → RPS → Storage math
`high-level-design.md`	Drawing the right boxes
`deep-dive-detailed-design.md`	Schema, APIs, data models
`database-scaling.md`	Read replicas, sharding, CQRS
`caching.md`	Cache layers, strategies, Redis
`load-balancing.md`	Algorithms, health checks, sticky sessions
`message-queues.md`	Kafka vs SQS vs RabbitMQ
`microservices.md`	Decomposition, contracts, versioning
`consistency-patterns.md`	Strong vs eventual vs causal
`rate-limiting.md`	Token bucket, leaky bucket, sliding window
`network-protocols.md`	HTTP/2, gRPC, WebSocket, SSE
`security.md`	AuthN/AuthZ, OAuth 2.0, JWT, TLS
`monitoring-observability.md`	Metrics, logs, traces — the 3 pillars
`storage-systems.md`	Object, block, file — S3, EBS, NFS
`databases.md`	Full DB selection guide
`trade-offs.md`	The core system design mindset
`wrap-up.md`	How to end an interview strongly
`insiders-guide.md`	What interviewers actually look for

🔁 The System Design Thinking Loop

Every good system design follows this loop. Internalize it:

🛠️ Local Setup — Get Running in 3 Minutes

bash

# 1. Clone the repository
git clone https://github.com/SOURAV-ROY/sdm.git
cd sdm

# 2. Install dependencies
npm install

# 3. Start the development server
npm run dev
# → Open http://localhost:5173

# 4. Build for production
npm run build

# 5. Format all files
npm run format

Action	How
Search any topic	Press `Ctrl + K` (or `Cmd + K` on Mac)
Jump to a section	Use the right-side On This Page outline
Browse all modules	Use the left Sidebar — all levels listed
See the full roadmap	Visit roadmap.md
Browse on GitHub	github.com/SOURAV-ROY/sdm
Live hosted site	souravsdm.vercel.app

🧠 Study Tips That Actually Work

✅ Do This	❌ Avoid This
Draw diagrams from memory after reading	Highlighting without thinking
Design each system 3+ times improving each pass	Reading once and moving on
Use the Feynman Technique — explain it out loud	Memorizing definitions
Keep a 1-page design journal per system	Skipping estimations
Time-box: 5min clarify → 10min HLD → 15min deep dive	Starting with solutions before requirements
Record yourself for 45min mock sessions	Designing in isolation without failure cases

You now have the full map. Pick your path and start.

🚀 Level 1 — Foundations · 🎯 Level 12 — Interview Prep · 🤖 LLMs & AI Agents

🌐 Live Site · ⭐ Star on GitHub

Made with ❤️ by Sourav Roy

🗺️ System Design Mastery — Complete Guide Life Flow ​

🌐 Big Picture — What Is This Repository? ​

🔭 Full Curriculum Overview (All 12 Levels) ​

🎯 Choose Your Learning Path ​

🟢 Path A — The Beginner (0–2 years experience) ​

🟡 Path B — The Intermediate Engineer (2–5 years) ​

🔴 Path C — The Interview Candidate (2–3 weeks sprint) ​

🟣 Path D — The Tech Lead / Architect ​

🤖 Path E — The AI / ML Engineer ​

📦 What's Inside Each Level ​

Level 1 — Foundations ​

Level 2 — Scalability ​

Level 3 — Databases & AI Data Systems ​

Level 4 — Caching ​

Level 5 — Messaging & Queues ​

Level 6 — Microservices ​

Level 7 — Real-World Systems (14 deep dives) ​

Level 8 — Case Studies & Interview Prep ​

Level 9 — Design Patterns ​

Level 10 — Monitoring & Observability ​

Level 11 — Academic Research ​

Level 12 — Interview Prep (20 deep-dive topic guides) ​

🔁 The System Design Thinking Loop ​

🛠️ Local Setup — Get Running in 3 Minutes ​

⌨️ Navigation Tips ​

🧠 Study Tips That Actually Work ​