The first thing we read is not your CV.
We read the part of the application where you describe something you actually built or ran, and what broke. That is the whole test. Most applications reach it in the last paragraph, after 300 words about ambition.
What evidence looks like
A credential is a receipt for having been in a room. For 30 years that receipt stood in for proof, and the employer paid for the potential they inferred from it. That arrangement has been repriced in public, and we put the numbers in why your degree was priced on scarcity. Knowledge stopped being rare. The receipt stopped meaning much.
Evidence is different, and it has a shape. It is a thing that ran, for a real user, with a failure mode you can describe out loud without rehearsing it. A scraper that fell over on page four and taught you to handle rate limits. A model that overfit, and the two weeks you spent finding out why. A market you cold-called until somebody said yes.
If you cannot describe something that broke, you have not shipped anything yet. That is fine. It is just a different conversation.
Why we cannot measure potential
We are not screening for promise. Promise is not measurable from a written form, and we do not pretend otherwise. What we can read is what has already happened. That gap is why we ask for the specific and not the aspirational.
The second reason is structural. The skill this programme tests sits at the intersection of three literacies almost no programme trains together: enough technical depth to argue with an engineer, enough financial depth to argue with a credit committee, and enough operational depth to run a site where the standard playbook does not apply. No lecture hall trains that. No transcript screens for it.
What we actually read
Numbers first. What you did, for how long, with what result, and how you know the result. Specifics survive contact with the work. Generalities do not.
Then we look for the description of a failure. An application that contains only successes reads as a brochure. We have read enough brochures. The people who have done the work lead with the thing that did not go to plan, because that is where the learning is and it is also the part they are still annoyed about.
Then we look for whether the work had a user other than you. A project graded by a teacher, a blog post read by nobody, and a tool that 40 people open every week are three different claims. We can tell them apart in one line.
25 out of 25,000
The programme takes 25 people per cohort from a pool that has passed 25,000 applicants. That is one in a thousand. Participants come from 26 nations and work across 15 live projects inside real companies in the group, with no grades, no exams and no syllabus. 210 alumni now work across 19 countries.
The ratio is not a marketing device. It is the number that decides how much of the programme is delivered by the practitioners running the ventures rather than by a teaching staff, because a practitioner's afternoon is expensive and there is only one of them. It is also why a placement runs four to six months on site and not in a classroom.
Current openings sit across six tracks: consulting, artificial intelligence and automation, engineering, marketing, sales, and EU funding and research. The board of open positions moves weekly. That is what you would expect from a live organisation, and not from a school.
The honest part
This is wrong for you if you want a certificate, a syllabus, or a tutor who catches what you break. There is no safety net on a four-person venture. The first three weeks are uncomfortable for almost everybody. It is also wrong for you if you want a summer to fill.
It is right for you if you would rather own something small and real than watch something large. Nobody here will ask whether you are ready. The probability that a credential survives the first week of live work is close to zero, and the application is where we find out which side of that line you are on.
If the work is the point, the board is here, and the first thing we will read is what you shipped.