The flagship Cohorte program. Live mentors, written feedback every fortnight, and a working AI system in production by week nine. No theatre.
Not a demo. A real AI system deployed to real users, measured and signed off by week nine of twelve.
A scoping doc, a verification plan, and a governance runbook — the operating layer your team inherits.
Weekly rubrics from three senior mentors. The grade is the transcript we sign.
Twelve months of Engine Room membership and priority access to advisory slots after graduation.
The curriculum is public. Every week has reading, an exercise, a 1:1, and a rubric. The syllabus is versioned; each cohort ships a small improvement.
Two weeks before kickoff. You meet your lead mentor, scope your project, and read the three papers we don't teach without.
How to size a use case, write a measurement plan, and refuse the wrong ones. The prompt is never the bottleneck; we spend three weeks proving it.
The verification curriculum CM wrote after shipping 60+ enterprise systems. Conformal prediction, self-consistency, and the TrustGate framework applied to your own project.
Your system goes live this phase. Deployment, monitoring, a written runbook, and the first audit your compliance team will sign.
Governance layer, external review, demo day. You leave with a system, a document, and a mentor signature on each.
Part-time by design. Real work gets done between Tuesday kickoff and Thursday review. The rest is reading, writing, and shipping. No 9-to-5 simulation.
A short Monday note from CM. One idea, one frame, one nudge.
Ten minutes of prep. The session assumes you've read the paper.
Two hours, cohort-wide. One topic, one case, one pushback.
One paper, one letter, one playbook chapter.
The week's applied exercise. Usually on your own project.
Post in cohort Slack. Read two peers. Leave one useful comment.
One of your three mentors. Unstructured. Whatever you're stuck on.
Written answer, not a slide. Mentor returns rubric within 48h.
Drop-in working session. CM or one mentor is always in the room.
Written feedback on your exercise. Two paragraphs, not a score.
Founded Cohorte in 2023 after a decade shipping AI systems in banks, insurance, and the public sector. Author of The AI OS. Runs the Foundations and Verification phases.
Runs production AI at a European fintech with 4M users. Sat where you'll sit. Leads the Shipping phase, runs the weekly 1:1s for half the cohort.
Writes the AI governance standards for one of Europe's top-ten banks. The only mentor you'll meet who reads every rubric before you submit it. Leads the Governance phase.
Scaling AI across a firm like PwC is where most organizations stall. Plenty of ambition, not enough system. Charafeddine built and led our AI Factory: the engine that turned ideas into dozens of shipped projects and doubled our AI capabilities in four months.
Applications take twelve minutes. We answer every one, yes or no, within seven days. No sales call. No funnel.
Apply to Cohort N°07