About
The practitioner's guide
to AI in the data stack.
AI for the Stack exists because most AI resources are written for the wrong audience. ML researchers get the papers. Non-technical people get the news digests. The data engineer in the middle — already running dbt and Airflow, being asked to "add AI" to a system they're responsible for keeping reliable — gets almost nothing useful.
This newsletter fills that gap. Every issue is written from the perspective of someone who has integrated LLMs into real production workflows — not toy demos, not weekend side projects, but actual pipelines under deadline pressure where reliability and cost matter.
The format is two issues per week:
Tuesday — Deep Dive
One real-world use case. A concrete workflow you can follow. Every issue ships a working GitHub repo with copy-paste-ready code — not snippets, not pseudocode. If it doesn't work in a real pipeline, it doesn't get published.
Thursday — Roundup
Three to five items: the best tools available right now, the workflow setups worth stealing from the community, and what's actually working in production this week. Curated and opinionated — no filler, no SEO-bait listicles.
The standard here is simple: if something is overhyped, we say so. If a tool doesn't hold up under real workloads, we say so. If there's a better way to do something than the way the blog posts describe, we show it.
"Written by practitioners, for practitioners. Not journalists, not researchers, not AI influencers — people who run the pipelines you run."
If you're a data or analytics engineer being asked to bring AI into your stack and you want a practical, honest guide to what actually works — this newsletter is for you.
Join the newsletter
Two issues a week.
Both free.
Tuesday: a real-world workflow with working code. Thursday: the best tools and setups in the space right now.
For data engineers. Not beginners. Not news aggregators.
Free. No spam. Unsubscribe anytime.