Writing
Notes on backend engineering, legacy migration, and AI-assisted retrieval — part notebook, part tutorial.
A practical walkthrough of the small neural network behind my interactive digit-recognition playground — from Python training to local TypeScript inference.
NN/g named five traits that separate good site chatbots from frustrating ones. Here's what each one actually costs you in retrieval, guardrails, and escalation code.
One AI output has to satisfy an analyst, an executive, and an end user at once. Explainability isn't a model feature — it's something your system has to emit on purpose.
A Go-to-C transpiler reignited the 'Go is a better C' debate. Where the claim holds up, where it falls apart, and my take as a backend dev.
Why I'm starting a blog, and what I plan to write about — backend engineering, legacy migration, and AI-assisted retrieval.
Vector search feels magical until it confidently returns the wrong thing. Here's why combining it with keyword retrieval works better.