About this journal.
Simple Machine Learning is a small, opinionated journal of essays on the practice of machine learning. It exists because most explanations of ML are either too shallow to teach or too dressed-up to read. The essays here aim for the middle — plainspoken, honest about uncertainty, generous with examples.
Typography
Display type is set in Bricolage Grotesque by Mathieu Triay — a precise, slightly expressive sans with subtle character. Body copy is set in Newsreader by Production Type, optimized for long-form reading on screens. Technical metadata uses JetBrains Mono.
Stack
Built with Next.js using static export and deployed to Netlify. Posts are plain Markdown files with front-matter; no CMS, no database, no analytics by default. The whole site is a folder you can put under git.
Adding posts
Drop a Markdown file into content/posts/:
---
title: A short title
description: A standfirst, one sentence.
date: 2026-04-29
author: Your Name
tags: [linear-algebra, intuition]
---
Your essay starts here. Markdown works.
## A section
> A pull quote.
```python
def loss(y, yhat):
return ((y - yhat) ** 2).mean()
```
Or scaffold one quickly:
npm run new-post -- "A short title"Then npm run dev and the post appears on the front page automatically. Set draft: true in front-matter to keep it out of the listing.