Writing & Technical Notes

Pagination Is Not a Detail You Take Lightly

Designing a Timeline API: The Simplicity of Fan-out-on-Read

Designing a Timeline API: The Simplicity of Fan-out-on-Read.

From CRUD to Distributed Systems: Why I'm Building a Twitter Timeline Service (and Documenting Every Phase)

A breakdown of Twitter's fan-out architecture - and why optimizing for reads over writes was one of the most elegant tradeoffs in social media engineering.

Dead Letter Queues: Building Reliable Distributed Systems That Don't Lose Messages

Retries alone aren't enough. Learn how Dead Letter Queues isolate poison messages, preserve failures for debugging, and keep distributed systems reliable.

The Outbox Pattern: How I Stopped Losing Messages Between My Database and My Queue

A payment succeeds, the customer gets charged, but the confirmation email never arrives - because your process crashed between two lines of code. Here's the pattern that closes that gap, and why every production system eventually needs it.

Building a Real-Time Analytics System with Apache Kafka: Lessons from My Mini Uber Project

I wanted to understand how companies like Uber process millions of location updates every day. Instead of watching more tutorials, I decided to build a small event-driven backend from scratch using Apache Kafka, Socket.IO, and TypeScript.

How Netflix Stopped One Slow Microservice From Taking Down the Whole Company

The real story behind Hystrix, Netflix's circuit breaker library - why cascading failures kept taking down their platform, and the pattern they built to stop it.

How X (Formerly Twitter) Adopted Fan-out So Beautifully

A breakdown of Twitter's fan-out architecture - and why optimizing for reads over writes was one of the most elegant tradeoffs in social media engineering.

You Can't console.log Your Way Out of a Distributed System

What DDIA Chapter 1 taught me about partial failure - and why tracing tools like Zipkin and OpenTelemetry aren't over-engineering.

Designing a Scalable Distributed Notification System

A deep dive into building a production-ready notification system with RabbitMQ, Redis, and proper observability.