"This is the internal training manual I wish every candidate had before they sat in our interview room." — Senior Staff Engineer (Google/Meta/Netflix)
MAANG interviews in 2026 are not about memorizing diagrams. They are about Decision-Making under Constraints.
This repo is a complete, zero-gap curriculum designed to take you from Zero to Staff-Level Architect. It's built by engineers who have conducted over 500+ interviews and have designed systems you use every day.
- L4 (SDE II): Can stay structured and build a working happy path. (Score:
Hire) - L5 (Senior): Can justify every tradeoff with numbers and explain "The Why." (Score:
Strong Hire) - L6+ (Staff): Can design for global failure, cost, and organizational scale. (Score:
Exceptional Hire)
Starting from scratch? Follow these in order. Already a Senior? Jump to Unit 06.
maang-system-design-playbook/
│
├── 00-Orientation/ # LEVEL CHECK: Where do you stand?
├── 01-Interview-Skills/ # STRATEGY: Whiteboard, Speech, Signals
│
├── 📚 THE FOUNDATIONS (Must Know)
├── 02-Deep-Fundamentals/ # LSM Trees, Raft, BGP, Clock Skew
├── 03-Building-Blocks/ # Redis, Kafka, Cassandra, LB, S3
├── 04-Patterns-and-Paradigms/ # Sagas, CQRS, Circuit Breakers, Outbox
├── 05-Decision-Tradeoffs/ # SQL vs NoSQL, CAP, Consistency vs Latency
├── 06-API-and-Data-Modeling/ # The Schema Design Masterclass (NEW)
│
├── 📐 THE CASE STUDIES (Practice)
├── 07-Case-Studies/ # Twitter, YouTube, Uber, Uber, Job Scheduler
├── 08-Diagrams/ # Mermaid Templates & Visual Strategy
│
├── 🚀 THE STAFF-LEVEL PLAYBOOK (L6+)
├── 09-Advanced-Topics/ # Multi-Region, Disaster Recovery, Chaos Eng
├── 10-Real-World-Failures/ # Cache Stampedes, Thundering Herds
├── 11-Scaling-Playbooks/ # Read-Heavy vs Write-Heavy systems
├── 12-Observability-and-SRE/ # SLOs, SLIs, Alerting for Architects
├── 13-Security-and-Compliance/ # GDPR, OAuth2, Data Residency
│
├── 🏢 THE COMPANY PATTERNS
├── 14-Company-Specific/ # How Google vs. Amazon vs. Meta build
│
├── 🤖 AI & FUTURE DESIGN
├── 15-AI-System-Design/ # LLM Backends, Vector DBs, RAG (2026 Hot Topic)
│
└── 🏁 FINAL PREPARATION
└── 16-Interview-Simulation/ # Mock Rubrics and Follow-up Questions
- Read the AUDIT-AND-TRANSFORMATION-PLAN.md: Understand what most candidates get wrong.
- Follow the ROADMAP-12-WEEKS.md: A day-by-day plan for deep prep.
- Self-Evaluate using PRACTICE-CHECKLIST.md: Know your level (L5 vs L6).
- Keep the QUICK-REFERENCE.md open during all mock interviews.
What we actually write down while you're talking:
| Signal | What we look for |
|---|---|
| Clarification | Does the candidate ask about "Out of Scope" features? |
| Calculations | Can they estimate storage for 1B users in their head? |
| Tradeoffs | "I use X for Y, accepting Z as a penalty." (Standard of Excellence) |
| Failures | "What happens when the Redis node dies?" (Zero to Hero) |
| Pacing | Do they finish the high-level design in < 15 minutes? |
Large Language Model (LLM) system design is now a standard question at MAANG (L5+).
- Unit 15 covers: Vector Databases, RAG (Retrieval Augmented Generation), and Prompt Caching strategies. Do not skip this.
Found a gap? While we have covered the Big 16 case studies, the web evolves. PRs that follow the New Template are always welcome. See CONTRIBUTING.md.
- Designing Data-Intensive Applications (Kleppmann) - The Bible
- System Design Interview (Volumes 1 & 2) (Alex Xu) - The Practical Guide
- Google SRE Book - For Staff+ Reliability Thinking
Built with ❤️ by Senior Staff Engineers who want you to succeed.