Blog
Thoughts, write-ups, and notes on building software, design, and everything in between.
Why LSTMs Work: The One Line That Makes Long-Range Memory Possible
August 24, 2026
Plain recurrent networks forget everything past about ten steps. LSTMs don't. The reason comes down to a single equation — and it's the same trick that makes ResNets and Transformers trainable.
API Gateway Deep Dive: Fundamentals, JWT Auth & Revocation Strategies
August 24, 2026
A comprehensive guide to how API Gateways work, how they validate JWTs, and the tradeoffs behind token revocation. Table of Contents Part 1: API Gateway Fundame...
Semantic Matching Without a Single Line of Training Data
August 15, 2026
How I designed an embeddingbased job–candidate matching service and why the interesting decisions were the ones I said "no" to. The setup I'm building a job pla...