Technical writer & docs engineer by day, Vibecoder by night.
I build the Bengal stack — a Python-native web stack for 3.14+ free-threading. Every layer is pure Python, typed, and designed for the modern runtime.
| ᓚᘏᗢ | Bengal | Static site generator — notebooks, autodoc, incremental builds |
| ⌁⌁ | Chirp | Web framework — HTML over the wire, htmx, streaming |
| =^..^= | Pounce | ASGI server — free-threading native, HTTP/2, WebSocket |
| )彡 | Kida | Template engine — Jinja2-compatible, AST-native |
| ฅᨐฅ | Patitas | Markdown parser — typed AST, ReDoS-safe, CommonMark |
| ⌾⌾⌾ | Rosettes | Syntax highlighter — lock-free, 55+ languages |
Python-native. Free-threading ready. No npm required.
I write and engineer documentation for developer products. I've shipped docs at:
| NVIDIA | AI (NeMo) |
| HPE | Enterprise AI Solutions |
| Pachyderm | Data pipelines |
| Bolt Financial | Fintech |
| Litify | Legal Tech |
| Metricly | Cloud monitoring |
| ncino | Banking platform |
I've used and migrated between just about everything: Salesforce Knowledge, Zendesk Guide, Hugo (golang), Sphinx (RST/Myst), MKdocs, Docusaurus, and Fern (MDX). I believe that documentation is both a product in its own right and fundamentally data at its core. It needs strategy and planning and can suffer from "tech debt" like any other project when corners are cut.