What are the three engineering builds?

A Technical Documentation Writer (turns code + design docs into shippable internal docs), an Incident Summarizer (post-mortems and on-call shift reports from logs + Slack threads), and a Technical Research Agent (vendor comparisons, library evaluations, RFC drafts). The capstone wires them into your engineering OS.

Is this the Engineering Series (E1–E6) or different?

Different. AI OS for Engineers (€297) is the operator-layer course: three high-leverage builds you use in your day-to-day eng work. The Engineering Series (€797 bundle) is the systems-architecture track — building an Enterprise AI OS from scratch. Start here, then take E1–E6 if you want to architect AI platforms.

I already use Copilot / Cursor / Claude Code. What's new here?

Copilot/Cursor write code. This course is about the engineering work that surrounds code: docs, incident reviews, research, architecture decisions. The judgment work that doesn't show up in your IDE.

Does this go into model internals, fine-tuning, or stay at the application layer?

Application layer. We use frontier models as-is, focus on prompt architecture, structured outputs, and tool use. Model-internals work is in the Engineering Series (E2 and beyond).

Claude or ChatGPT paid plan. Your existing dev environment. Optional: an API key for the more advanced builds. No new infrastructure required.

8–10 hours: ~5 hours of lessons, ~5 hours building. Plan one build per evening across a week.

Can my company pay for this?

Yes. Engineering managers commonly approve this as professional development. Invoices issued. Email hello@cohorte.co subject 'Reimbursement.'

What's the refund policy?

€297 courses are non-refundable once you open the course. Unopened, they are refundable in full within 14 days of purchase. Courses above €297 stay refundable after you open them, as long as you have not completed the first module. This one sits at €297, so that window does not apply.

For Engineers

Code that compiles
is not code that ships.

Three builds that keep your engineering judgment in the loop. Tech Doc Writer. Incident Summarizer. Technical Research Agent. AI handles the routine. You own the architecture.

Get on the waitlist

€297 waitlist

Self-paced

The problem

The auth module stored tokens in localStorage.
Policy required HttpOnly cookies.

An AI-generated authentication module stored session tokens in localStorage. Company policy required HttpOnly cookies. Nobody caught it. For six weeks.

By the time someone noticed, the pattern had been copied into three other modules. The fix took two sprints. The security review took another.

The code compiled. The tests passed. The deployment succeeded. The AI did exactly what was asked.

The problem was that nobody asked the right question: does this match our security policy?

AI did exactly what was asked. Nobody asked the right question.

AI doesn't read your company's architecture decision records. It doesn't know your team's conventions. It doesn't know that your CTO banned localStorage for tokens in a Slack message three months ago.

The cost of a wrong engineering decision compounds through your entire system. AI makes that compounding faster.

This course teaches you how to build AI systems that respect the engineering judgment boundary, accelerating the routine while protecting the decisions that matter.

The method

13 lessons. 5 modules.
One engineering OS.

AI handles the routine. You own the architecture. The mind stays in the loop where it counts.

Module 1 · Primer 01

Where AI belongs in engineering

The Engineering Judgment Boundary

Why code that compiles is not code that ships. The localStorage incident dissected. Where AI accelerates engineering work and where human judgment is non-negotiable.

2 lessons

Module 2 · Build 1 02

Documentation that ships

Technical Documentation Writer

Upload existing docs as style examples. Define format: API docs, runbooks, ADRs. Bake in your team's conventions. Test, iterate, ship.

3 lessons

Module 3 · Build 2 03

Post-mortems that travel

Incident Summarizer

Structured post-mortem format: timeline, root cause, impact, resolution, action items. Feed raw logs, Slack threads, ticket comments. Get a structured doc you can actually share.

3 lessons

Module 4 · Build 3 04

Research with rigor

Technical Research Agent

Research agent searches benchmarks, comparisons, community feedback. Challenger agent audits for bias and missing alternatives. Manager synthesizes. You decide.

3 lessons

Module 5 05

Run it weekly

Your Engineering OS

Connect the three builds. Engineering-specific verification protocols. Code review with AI as assistant, not authority. Capstone: ship one real workflow end to end.

2 lessons + capstone

The three builds

Three transformations. One engineering OS.

Each build replaces a real engineering bottleneck. AI handles the routine. You stay in the loop where the architecture lives.

Tech Doc Writer

From draft to done.

45-minute docs → 10-minute review

A specialist bot that generates documentation in your team's style. API docs, runbooks, ADRs. Consistent format. Every time.

Incident Summarizer

From logs to lessons.

2-hour post-mortem → 15-minute review

Raw incident logs, Slack threads, and ticket comments become structured post-mortems. Timeline, root cause, impact, resolution, action items.

Technical Research Agent

From browsing to deciding.

Full-day research → 30-minute review

Research and Challenger agent team. Evaluates frameworks, tools, approaches. Audits for bias and missing alternatives. Manager synthesizes.

Honestly

This is for you if:

→You're a software engineer, data scientist, technical lead, or CTO

→You use AI for coding and want a review system that catches policy violations

→You spend hours on documentation and post-mortems

→You evaluate tools and frameworks and want faster, more rigorous research

→You want AI to accelerate the routine without compromising architecture

Don't take this if:

→You want to build production AI systems. See the Engineering Series E1-E6.

→You want AI to write your code without review. That's the trap.

→You're looking for non-technical AI skills. See the Thinking Stack.

Pricing

One price. Lifetime access.

€297

One-time payment. All future updates included.

13 lessons across 5 modules (video and written)
Technical Documentation Writer build guide and templates
Incident Summarizer build guide
Technical Research Agent build guide
ADR template, code review checklist, and prompt vault

3 months in the Engine Room. Where alumni and operators go to get unstuck.

Get on the waitlist

FAQ

Before you ask.

Anything else? Email [email protected].

What are the three engineering builds?: A Technical Documentation Writer (turns code + design docs into shippable internal docs), an Incident Summarizer (post-mortems and on-call shift reports from logs + Slack threads), and a Technical Research Agent (vendor comparisons, library evaluations, RFC drafts). The capstone wires them into your engineering OS.
Is this the Engineering Series (E1–E6) or different?: Different. AI OS for Engineers (€297) is the operator-layer course: three high-leverage builds you use in your day-to-day eng work. The Engineering Series (€797 bundle) is the systems-architecture track — building an Enterprise AI OS from scratch. Start here, then take E1–E6 if you want to architect AI platforms.
I already use Copilot / Cursor / Claude Code. What's new here?: Copilot/Cursor write code. This course is about the engineering work that surrounds code: docs, incident reviews, research, architecture decisions. The judgment work that doesn't show up in your IDE.
Does this go into model internals, fine-tuning, or stay at the application layer?: Application layer. We use frontier models as-is, focus on prompt architecture, structured outputs, and tool use. Model-internals work is in the Engineering Series (E2 and beyond).
Stack requirements?: Claude or ChatGPT paid plan. Your existing dev environment. Optional: an API key for the more advanced builds. No new infrastructure required.
Time commitment?: 8–10 hours: ~5 hours of lessons, ~5 hours building. Plan one build per evening across a week.
Can my company pay for this?: Yes. Engineering managers commonly approve this as professional development. Invoices issued. Email [email protected] subject 'Reimbursement.'
What's the refund policy?: €297 courses are non-refundable once you open the course. Unopened, they are refundable in full within 14 days of purchase. Courses above €297 stay refundable after you open them, as long as you have not completed the first module. This one sits at €297, so that window does not apply.