01 — BASELINE
Measure first
Start with a bounded diagnostic and baseline tasks under stated AI-access conditions.
Human capability for work done with AI
Your next profession is aHuman–Agent Work Cell
Not a course. Not a certificate. Build a working system of AI agents you direct — turning what you already know into real, inspectable work and showing where your judgment changes the result.
Early research-stage project. No employment, certification or income guarantees.
NOW
AI
NEXT
capability → supervision → evidence
The problem
The value of what you know is falling. The value of what you can be trusted with is rising
AI makes explanation, drafting and first-pass execution increasingly abundant. That does not make human judgment irrelevant. It makes verification, scope, correction and accountability more important.
REPROF is testing a simple idea: can people build new professional capability faster with AI — while producing evidence that shows not only what they made, but how they supervised the machine?
We treat that as a hypothesis to measure, not a promise to sell.
The REPROF method
Capability, not content
The program, tutor and community are mechanisms. The output is observable work, independent scoring and a record of human judgment over AI.
02 — PRACTICE
Do real work
Thirty-day profession tracks are organised around graded work products, not lesson consumption.
03 — SUPERVISE
Catch the machine
We record where the human corrected, rejected, stopped or overruled AI — and where they missed an intervention.
04 — EXTERNALISE
Use outside rulers
Third-party lab scores and blind external grading make the evidence more credible than a self-issued badge.
05 — COMPARE
Measure the delta
Day 0 and Day 30 are compared under equivalent conditions, including confidence and error-control.
06 — OWN THE PROOF
Leave with evidence
Artifacts belong to the participant: work sample, capstone, external scores and supervision record.
“AI literacy is easy to claim. Supervision is harder to fake.”
REPROF founding principleThe evidence stack
Proof a stranger can inspect
External lab resultsMachine-verified completion on established platforms.
Blind-graded capstoneA realistic work artifact scored against a public rubric.
AI Supervision RecordWhere the human corrected, rejected, stopped or overruled AI.
Capability DeltaWhat changed between comparable baseline and endpoint conditions.
Profession #01
AI-Assisted Cybersecurity Researcher
Cybersecurity is our first research domain because performance can be externally scored, false positives are observable, and professional boundaries are explicit.
FOUNDING RESEARCH TRACK
Thirty days. Authorised labs. A documented capability ceiling
The intended Day-30 outcome is bounded: independently perform a scoped assessment of a web application in an authorised lab, verify common vulnerability classes, control false positives, explain severity and remediation, and document where AI was corrected or rejected.
What this does not mean: completion does not make someone a penetration tester, does not authorise testing real systems, does not create an accredited certification and does not guarantee employment.
- External lab scoring
- AI-assisted analysis with human accountability
- False-positive control
- Professional reporting
- Scope and authorisation discipline
- Capstone + supervision record
First commercial offer
Founding Cohort #01 is now open for applications
A 30-day capability-and-evidence sprint for AI-Assisted Cybersecurity Researchers. Apply free; accepted applicants begin through a private €49 Entry Activation.
Founding price€490See the offer →
Private commission
Human–Agent Work Cell Foundry
Turn existing expertise into a governed Sense → Deliver → Prove operating system — privately built, tested and handed over in 90 days.
Application-only build€4,900Explore the Foundry →
Research before marketing
We measure what changed. Then we publish it — including when it doesn’t
REPROF will pre-register its hypotheses, publish its rubric, report sample size and limitations, and include dropout, null and negative results.
Free front door
What can you actually do with AI?
The REPROF Capability Diagnostic is a short assessment of verification discipline, error detection, recovery, scope, communication and orchestration. It is not a personality test, certification or job prediction.
Take the free diagnostic →✓
Immediate readout
See what you demonstrated under defined conditions.
See what you demonstrated under defined conditions.
!
Supervision risks
See where you may be trusting confident AI output too quickly.
See where you may be trusting confident AI output too quickly.
↗
Next practice
Get a concrete next action, not a career promise.
Get a concrete next action, not a career promise.
Who is behind this
Transparent before impressive
REPROF is a new research-stage project founded by Steven Alber. The founder is not presented as an elite cybersecurity practitioner. Profession-specific technical claims require named external technical authority and independent grading before they are published as evidence.
That distinction is deliberate: strategy, research architecture and product design are separate from domain-specific technical authority.
Read the method and limitsFounding phase
Follow the experiment before the outcome is known
Cohort Zero is a research input, not proof. Global development continues independently; public claims wait for data.