Contribute.
We publish articles, case studies, primers, and tools about ML systems — from anyone learning or working in the space. If you've figured something out and want to share it with the world, you should write about it here.
The process
Share an idea
Open a thread in our Blog ideas discussion on GitHub. A few sentences is enough — we will take it from there.
Shape it
We'll help you narrow the idea, find the useful angle, and decide what form it should take.
Draft
Write it your way. Use our in-browser editor at /write — no markdown needed — or write MDX directly if you prefer. Either way, add code, figures, equations, or links if they help.
Review
We'll review for clarity, correctness, and readability. Nothing fancy, just make the work easier to learn from.
Publish
Once it is ready, we publish it and credit you as a contributor. Comments happen in GitHub Discussions.
Two ways to write
New to this? Use the in-browser editor — add headings, images, tables, video, and math visually, then download a ready-to-publish folder. No markdown, no setup.
Comfortable with MDX? Write it directly. Either way, send it to us as a pull request or by email and we'll take it from there.
What we publish
- Articles — opinionated, well-argued takes on a real problem (~1,500–3,000 words).
- Case studies — what happened when you actually shipped the thing.
- Primers — careful explainers of techniques that deserve more careful treatment.
- Tools — working calculators or visualizers, with a writeup.
- Reading lists — annotated bibliographies on narrow topics.
What we don't publish
- Promotional content for products.
- Recycled paper summaries.
- "X is dead, long live Y" hot takes without evidence.
- Listicles. Especially listicles.
Style
Concrete over abstract. Numbers over adjectives. One idea per paragraph. Show your work, including the parts that didn't.
We're trying to make a publication that you'd want to send to your future self. Write accordingly.