The Complete System Design Playbook with Ashwini Harle

Most system design resources hand you a handful of case studies to read once, or lock real practice behind a subscription. This is different: 60 fully worked problems with lifetime access, each one built as a complete, drillable mini-course, not a summary. You don't just read the answer, you work through the journey, the estimation, the deep dives, the mistakes, and drill it until the reasoning is reflex.

Why Most Prep Doesn't Actually Work

If you're starting from zero, most resources assume you already know what a load balancer does or why everyone keeps talking about CAP theorem, which makes the whole subject feel intimidating before you've even begun.

And if you already know the theory, the problem is different: nobody taught you how to turn that knowledge into a complete, defensible design, out loud, on a clock, while someone watches you think. Courses hand you theory. Interviews test execution.

I'm a MAANG SDE-2. This playbook is the exact method and problem set I built from what actually earns a "yes" in the room, starting from the fundamentals and building all the way up to the hardest problems asked at top companies.

What's Inside?

PART 1: High-Level Design, Foundations → 30 Real Systems

36 chapters that start from scratch and teach each concept exactly when you'd reach for it in an interview:

1. Introduction to system design: what scalability, latency, and throughput actually mean
2. The HLD method: breadth-first then depth, on a clock (requirements → estimation → APIs → boxes → deep dives)
3. NFRs, availability and the nines, latency vs throughput (p50/p99)
4. CAP Theorem, PACELC, consistency and quorum
5. Rate limiting, DNS, API design (REST, RPC, GraphQL)
6. Sharding, consistent hashing, caching, indexing (B-tree vs LSM)
7. Replication and schema design
8. Database internals from scratch: WAL, LSM-trees, SSTables, Cassandra/DynamoDB
9. Search fundamentals: tries, inverted index, TF-IDF
10. Kafka, the AWS data stack, cloud essentials, back-of-the-envelope estimation

Then you put it to work across 30 systems designed end-to-end, each one a complete mini-course: journey walkthrough, estimation, deep dives tied to the exact requirement forcing that decision, common mistakes, "what if" variations, and flashcards for review. Systems covered include URL Shortener, Snowflake ID, Uber, WhatsApp, Netflix, News Feed, Dropbox, Instagram, Google Maps, Web Crawler, Ad Click Aggregator, Google Docs, Google Meet, Distributed KV Store, Metrics and Monitoring, a ChatGPT-style LLM service, and more.

PART 2: Low-Level Design, The Round That Blindsides Everyone

13 foundation chapters, starting from first principles and built around one repeatable method: clarify → entities → class diagram → code skeleton → pivots.

1. Object-oriented basics, SOLID principles, UML and class diagrams
2. Coupling, cohesion, dependency injection, error handling
3. All 22 design patterns, deep-dived and actually applied, never just name-dropped
4. Concurrency: locks, race conditions, deadlocks
5. A fully narrated mock interview, plus a red-flags catalog of what gets candidates rejected

Then 30 practice problems, each a full mini-course: Parking Lot, Vending Machine, Elevator, ATM, LRU/LFU Cache, Splitwise, Movie Ticket Booking, Payment Gateway, Digital Wallet, Pub/Sub, Ride Booking, Food Delivery, Chess, a Stock Exchange Matching Engine, and more.

PART 3: Cheat Sheet + Pattern Index

A one-page cheat sheet for the morning of your interview, plus a Pattern → Problems index so you instantly know which pattern fits which question.

What Changes for You After This

• If you're starting from zero, you'll go from not knowing what a system design interview even looks like to confidently running one, start to finish
• You can run a complete HLD in 45 minutes without losing the thread or running out of time
• You can turn vague requirements into a clean, extensible object model, and defend every class you draw
• You can choose databases, caching, sharding, and consistency models using real trade-offs, not memorized buzzwords
• You walk into the machine-coding round with a method instead of dread
• You narrate your thinking and adapt live, which is the actual thing interviewers are scoring

Who This Is For

• Complete beginners who've never designed a system before and don't know where to start
• SDE-1 candidates facing design rounds for the first time
• SDE-2 / SDE-3 engineers targeting senior offers at top product companies
• Backend and full-stack engineers leveling up or switching jobs
• Final-year students and new grads who want a method, not just answers
• Anyone who can recite the theory but freezes the moment they have to actually build something

Why This Is Different

• 60 worked problems you own for life, not a handful of case studies to read once, and not a subscription you lose access to
• Every problem is a drillable mini-course: journey walkthrough, estimation, deep dives tied to the exact requirement forcing them, common mistakes, "what if" variations, and flashcards, not just a worked solution and a diagram
• Starts from zero and goes all the way to hard interview problems, so you don't need a separate beginner resource before this one
• Goes first-principles where most courses stop short: build a database from WAL/LSM/SSTables, design an actual matching engine
• Written by a practicing MAANG SDE-2 solving these problems for real, not repackaging someone else's course

Others give you a handful of case studies to read, or a subscription to rent. This gives you 60 worked problems you keep forever, each one a mini-course you drill until it's reflex.