Local Tesla AI
A fully local, in-browser AI assistant built with Astro and WebLLM. Answers about Nikola Tesla from a local knowledge base — no backend, no data leaves your machine.
- @mlc-ai/web-llm
- astro
- dompurify
- fuse.js
- highlight.js
- marked
Production AI, not the demo kind.
I build RAG systems, agents and LLM tooling that hold up in front of real users — and I write down what actually survives. Behind it sit 18+ years of shipping software. This is where I keep the work: the notes, the experiments, and the projects that taught the lessons.
How I think about the work
Good software is less about the newest stack and more about the right one. I reach for the smallest set of tools that solves the problem, and I try to leave every codebase a little easier to change than I found it.
These days that means production AI: retrieval, LLM integrations and agents that survive contact with real users. I build them on nearly two decades of full-stack and architecture work — and what holds up in real use is exactly what I write about here.