3. HLD Interview Problems

27 curated system design problems organized by difficulty

How to Use This Catalog

1. Start with Easy: Build confidence with fundamental concepts

2. Progress to Medium: Apply advanced patterns (caching, sharding, replication)

3. Tackle Hard: Complex distributed systems, multi-region, consistency challenges

For Each Problem:

• Use hld-template.md for structured approach

• Practice capacity estimation with capacity-estimation.md

• Reference trade-offs-cheat-sheet.md for decisions

---

Difficulty Levels

Easy (1-2 weeks experience)

Focus: Basic architecture, simple scaling, single DB

Medium (SDE-2 level)

Focus: Distributed systems, caching, replication, load balancing

Hard (SDE-3/Staff level)

Focus: Multi-region, consensus, complex trade-offs, production operations

---

Problem List

Easy Problems (7 problems)

#	Problem	Key Concepts	Estimated Time
1	URL Shortener	Hashing, Base62, DB sharding	45 min
2	Pastebin	Key generation, S3 storage, CDN	45 min
3	Key-Value Store	Hash map, persistence, simple replication	45 min
4	Rate Limiter	Token bucket, sliding window, Redis	45 min
5	Web Crawler	BFS/DFS, politeness policy, distributed queue	60 min
6	Autocomplete	Trie data structure, caching, prefix search	45 min
7	Unique ID Generator	Snowflake, UUID, timestamp-based	30 min

Interview Frequency: ⭐⭐⭐⭐⭐ (Very Common)

---

Medium Problems (12 problems)

#	Problem	Key Concepts	Estimated Time
8	Twitter	Timeline generation, fan-out, Redis caching	60 min
9	Instagram	Image storage (S3), CDN, follower/feed model	60 min
10	YouTube	Video transcoding, CDN, recommendation algorithm	60 min
11	Notification Service	Push notifications, pub-sub, fan-out	60 min
12	News Feed	Ranking algorithm, pagination, denormalization	60 min
13	Chat System (WhatsApp)	WebSockets, message queue, read receipts	60 min
14	Search Engine (Google)	Inverted index, PageRank, distributed crawling	75 min
15	Dropbox	File chunking, sync, version control, delta sync	60 min
16	Netflix	Video streaming, adaptive bitrate, CDN	60 min
17	Uber	Real-time location, geospatial indexing (quadtree)	60 min
18	Yelp/Nearby	Geohashing, proximity search, read-heavy	60 min
19	API Rate Limiter	Token bucket, distributed rate limiting, Redis	45 min

Interview Frequency: ⭐⭐⭐⭐ (Common)

---

Hard Problems (8 problems)

#	Problem	Key Concepts	Estimated Time
20	Distributed Message Queue (Kafka)	Partitioning, replication, consumer groups	75 min
21	Payment System (Stripe)	ACID transactions, double-entry bookkeeping, idempotency	75 min
22	Stock Exchange	Order matching, low-latency, consistency	75 min
23	Ticketmaster (Booking)	Concurrency, distributed locking, overbooking prevention	75 min
24	Google Drive	Conflict resolution, operational transform, sync	75 min
25	Distributed Cache (Redis Cluster)	Consistent hashing, replication, failover	60 min
26	Search autocomplete (Google-scale)	Trie sharding, distributed aggregation, caching	60 min
27	Ad Click Aggregator	Stream processing, Lambda architecture, late data handling	75 min

Interview Frequency: ⭐⭐⭐ (Occasional, Senior+ roles)

---

Problem Selection Guide

For FAANG Interviews

Most Common (Practice First):

1. URL Shortener (Easy)

2. Rate Limiter (Easy)

3. Twitter/News Feed (Medium)

4. Notification Service (Medium)

5. Chat System (Medium)

Company-Specific:

• Google: Search Engine, YouTube, Autocomplete

• Meta: News Feed, Instagram, Chat System

• Amazon: Product Search,Distributed Cache

• Netflix: Video Streaming

• Uber: Real-time Location, Geospatial Search

---

Practice Roadmap

Week 1-2: Easy Problems

• URL Shortener

• Pastebin

• Rate Limiter

• Unique ID Generator

Goal: Nail fundamentals (API design, capacity estimation, basic scaling)

Week 3-4: Medium Problems (Part 1)

• Twitter

• Notification Service

• Chat System

• Instagram

Goal: Master caching, replication, fan-out patterns

Week 5-6: Medium Problems (Part 2)

• Search Engine

• Netflix

• Uber

• Dropbox

Goal: Geospatial indexing, CDN, real-time systems

Week 7-8: Hard Problems

• Distributed Message Queue

• Payment System

• Stock Exchange

Goal: Consensus, transactions, extreme consistency/latency requirements

---

Key Patterns Across Problems

Pattern 1: Caching (80% of problems)

Problems: Twitter, Instagram, YouTube, Yelp, Autocomplete
Strategy: Cache-aside with TTL, Redis/Memcached
Key: Identify hot data (80/20 rule)

Pattern 2: Sharding/Partitioning (60% of problems)

Problems: URL Shortener, Twitter, Message Queue, Distributed Cache
Strategy: Consistent hashing, range-based, hash-based
Key: Choose partition key to distribute load evenly

Pattern 3: Pub-Sub/Message Queue (50% of problems)

Problems: Notification, Chat, News Feed, Kafka
Tools: Kafka, RabbitMQ, SQS
Key: Decouple producers/consumers, async processing

Pattern 4: Geospatial Indexing (10% of problems)

Problems: Uber, Yelp, Nearby
Data Structures: Quadtree, Geohash, S2 geometry
Key: Efficient proximity search

Pattern 5: CDN (40% of problems)

Problems: YouTube, Netflix, Instagram, Pastebin
Strategy: Push vs Pull CDN, cache static assets
Key: Serve from edge locations, reduce origin load

---

Resources

• Template: hld-template.md

• Estimation: capacity-estimation.md

• Trade-offs: trade-offs-cheat-sheet.md

• Numbers: numbers-to-know.md

---

Interview Tips

1. Time Management: 15 min estimation, 30 min design, 15 min deep dive

2. Start Simple: Single server → Scale out

3. Justify Choices: "I chose Redis because..." (not just "use Redis")

4. Ask Questions: Clarify requirements before diving in

5. Think Production: Monitoring, failures, costs

Good luck! 🚀