LLM-Driven System Design Practice

Vision

Traditional system design interview preparation often relies on passive learning - reading articles, watching YouTube videos, and memorizing architectural patterns. While valuable, this approach lacks the hands-on experience that comes from actually building systems.

This repository represents a different approach: learning system design by actually implementing it.

With the assistance of LLMs like Claude, we can rapidly prototype lightweight versions of classic (and creative) system design interview questions. This hands-on practice helps develop:

• Architectural thinking - Making real trade-offs, not theoretical ones
• Implementation experience - Understanding what actually works vs. what looks good on a whiteboard
• Technology selection - Choosing the right tools for each use case
• Scalability patterns - Building systems that can grow
• Problem-solving skills - Debugging and iterating on real implementations

System Design Challenges

Each folder contains an independent implementation of a system design problem.

URL Shortening & Web Services

Storage & File Systems

Social Media & Content

Messaging & Communication

Video & Streaming

E-Commerce & Marketplaces

Transportation & Logistics

Ticketing, Events & Reservations

Developer & Productivity Tools

Financial Services & Payments

Business Services

Search & Discovery

ML/AI & Data Platforms

Infrastructure & System Components

UI Components & Patterns

Personal Projects & Experiments

External projects demonstrating system design concepts (with local architecture documentation):

Project Structure

Each project folder contains:

Getting Started

1. Choose a system design challenge that interests you
2. Read the architecture.md to understand the design decisions
3. Follow the README.md for setup and implementation details
4. Review claude.md to see how the system evolved through LLM collaboration

LLM-Assisted Development

This repository leverages Claude and other LLMs to:

• Rapidly prototype architectural ideas
• Generate boilerplate code and infrastructure
• Explore different implementation approaches
• Debug and optimize systems
• Document design decisions

See CLAUDE.md for guidelines on collaborating with AI on system design projects.

Scripts

Utility scripts for repository management:

# Count SLOC for entire repository
node scripts/sloc.mjs

# Count SLOC for specific project
node scripts/sloc.mjs scale-ai

# Output as JSON
node scripts/sloc.mjs scale-ai --json

# Output summary for README embedding
node scripts/sloc.mjs scale-ai --summary

Learning Approach

1. Understand the requirements - What problem are we solving?
2. Design the architecture - What components do we need?
3. Identify key challenges - What are the hard parts?
4. Implement incrementally - Start simple, add complexity
5. Test and iterate - Does it work? Can it scale?
6. Reflect and document - What did we learn?

Skills Developed

• Distributed systems design
• Database schema design and optimization
• API design and versioning
• Caching strategies
• Load balancing and horizontal scaling
• Real-time data processing
• Message queues and event-driven architecture
• Security and authentication
• Monitoring and observability

Contributing

This is a personal learning repository, but feel free to fork it and create your own implementations! Different approaches to the same problem are valuable learning opportunities.

License

MIT License - Feel free to use this for your own learning and interview preparation.

---

Remember: The goal isn't to build production-ready systems, but to gain hands-on experience with system design concepts. Start simple, iterate, and learn by doing!