For Engineers · Track 01

Manage Agentic Engineers.

You can write the code. Learn to direct the team writing it with you. A coaching sprint built around Claude Code and Codex from the terminal.

01Who this is for

If you are

  • A mid-to-senior engineer comfortable shipping production code.
  • An engineer who has not had to make architecture decisions, manage other engineers, or design the process around how work flows.
  • Using, or about to use, Claude Code or Codex from the terminal. Not IDE assistants.
  • An engineer whose bottleneck has shifted from writing code to deciding what to write and verifying what came back.

Probably not you if

  • You are learning to code. This is not a bootcamp.
  • You only work in an IDE assistant (Cursor, Copilot, JetBrains AI). The track is terminal-first.
  • You want a certification. The deliverable is capability, not a credential.
  • You want someone to write code for you. This is coaching, not contracting.
02What's broken without it

The job changed. The training did not.

The thing AI changes is not how fast you can write code. It is the size of the responsibility surface around each engineer. You are now expected to scope, decompose, delegate, review, and integrate work that used to belong to a team. The skills required for that are not the skills that got you here, and most engineering careers never required them.

03The sprint · 4 weeks · 1:1

Four weeks. One operating system.

Week 1

Intake

Current Claude Code or Codex usage, the repo, the gaps. Architecture review of one project you own.

Week 2

Work breakdown

Turning fuzzy goals into specs an agent can execute and a reviewer can verify. Defining "done."

Week 3

Process

Multi-session context management, multi-agent coordination, drift recovery, review without reading every line.

Week 4

Your OS

Written playbook, repo conventions, escalation rules, retro. The artefact you keep.

What's covered.

  • Architecture thinking for AI-assisted codebases
  • Requirements decomposition and work breakdown
  • Choosing what to delegate vs what to keep
  • Defining "done" so an agent can hit it
  • Context management across long sessions
  • Multi-agent coordination patterns
  • Terminal-first workflow (Claude Code, Codex)
  • Reviewing AI output without reading every line
  • Recovery from drift and wrong-direction sessions
  • Repeatable spec / PR / commit process
  • Repo hygiene and reproducibility
  • Documentation and handoff
04Pricing

Same pricing across every track.

1 credit = 30 minutes of working time. $150 per credit.

Start here

Diagnostic

2credits
$300 · 90-min call

Diagnoses the work, writes you a plan of action. Credited toward your sprint.

Book Diagnostic
Teams 2–5

Small Team

60credits
$9,000 · 6 weeks · up to 5

An engineering team learning to direct AI work as a unit. Shared playbook, async channel.

Charter: pay 48 credits ($7,200), receive 60.
Apply
Bespoke

Premium Bespoke

140credits
from $21,000 · 8–10 weeks

Cross-functional engineering org, leadership cell, or higher-stakes rollout.

Charter: pay 112 credits ($16,800), receive 140.
Request Call

Prices in USD · ex. VAT · settled in CHF

05FAQ
Do I need to be a senior engineer?
Mid-level and up. You need to be comfortable shipping code; the work here is everything around it.
Does this cover Cursor or other IDE assistants?
No. The track is built around Claude Code and Codex from the terminal, which is how I work. The discipline transfers; the workflows do not.
Will this make me a manager?
It will make you the kind of engineer who can architect and direct work, with or without the title.
Will this work if I am not allowed to use AI at work?
Yes. We will work on your side projects and personal repos. The discipline transfers when policy changes.

Start with a Diagnostic.

90-min call, written plan, credited toward your sprint. I will tell you straight whether this is the right track for where you are.