Notes on building systems
that survive real traffic.
Engineering write-ups from my projects — queues, CRDTs, deploys, and the trade-offs behind them.
05 POSTS
0102030405
How I Built a Fault-Tolerant Notification System with BullMQ
A production-focused breakdown of idempotency, retries, batching, and real-time delivery in the NotifyX notification pipeline.
CRDT vs Operational Transform: What I Learned Building CollabDocs
Why collaborative editing breaks naive sync, and how Y.js CRDT solved concurrent editing without silent data loss.
What Shipping Real-Time Apps Taught Me About Production
A practical write-up of the production bugs behind DevConnect: reconnects, CORS, token expiry, env drift, and message ordering.
How I Made Curlix Redirects Fast Without Losing Analytics
The core tradeoff behind Curlix: keep redirects sub-10 ms while still collecting useful click analytics asynchronously.
Why I Queued the LLM Work in Paper Pilot
Paper generation takes seconds, not milliseconds, so Paper Pilot pushes LLM and PDF work into a background worker and streams progress to the user.