Fieldwork records what held and what broke; training teaches the method underneath it. Open programs anyone can follow free, with public repositories attached, and private workshops for teams that want the same material hands-on, taught by the people who run these systems in production.
LEARN THE METHODFieldwork records what held and what broke. Training teaches the method underneath it: open programs anyone can follow free, and private workshops for teams that want the same material hands-on.
Each open program is a numbered post with a public repository attached, MIT licensed, self-paced, no signup.
A free, open program that takes a motivated beginner from a first Python notebook to a governed production lakehouse, by way of retrieval, fine-tuned models, agents, three public clouds, and Databricks.
Each open program is a numbered post with a public repository: MIT licensed, self-paced, no signup, no paywall. The AI Engineering Lab carries a motivated beginner across 24 weeks to production discipline, and new programs append to the register as they publish.
A team, a room (or a call), your stack, and a Zorost engineer. Workshops adapt the open curriculum to your data constraints and current level: retrieval done right, fine-tuning with an evaluation gate, agents under command, Databricks with governance. On-site and air-gapped delivery available.
Teaching is the proof habit applied to knowledge: if the method only works when kept secret, it is not a method. The open programs are the same discipline we sell, stated plainly, so a buyer can inspect exactly what a workshop will build in their team.
The register grows with the Lab: fine-tuning practicals from the Qwen program, local AI operations from the Spark Duet, and agent engineering from the production series are the natural next entries. TRAINING / 02 is next, and the number is already reserved.
Two forms: open programs published free with public repositories (the 24-week AI Engineering Lab is the first), and private workshops where a team works the same material hands-on with a Zorost engineer, scoped to their stack and data.
Yes. MIT licensed, no signup, self-paced, and the cloud weeks are designed to stay inside free tiers. Seven phases, 43 runnable notebooks, one continuous case study, from first Python notebook to a governed production lakehouse.
The engineers who run these systems in production. The material comes from the Lab's own registers: the same retrieval, fine-tuning, agent, and Databricks discipline documented across the site, taught hands-on rather than read.
That is the default. A workshop starts from your stack, your data constraints, and your team's current level; the open programs provide the base curriculum, and the engagement adapts it. Air-gapped and on-site delivery are available.
We don't pitch slide decks. We show you what we've already built in your domain, then engineer what your mission requires.
