Think tank on the industrialization of AI in regulated systems

Industrial AI in critical, highly regulated environments.

The Twingital Institute is an independent research space dedicated to the real conditions of deploying artificial intelligence in critical and heavily regulated environments.

I don't build products.
I design the architectural, governance and compliance frameworks that allow AI to exist as a sustainable operational capability.

Healthcare & Life Sciences · Public Sector · Defence · Critical Infrastructure · Regulated Industries

Explore the RAISE Framework Access publications

—

Doctrine

Most of the debate around AI focuses on algorithmic performance — without having resolved how to measure, validate or standardise it in real operational contexts.

But even when proven, documented and validated (including the safety of the algorithm in use), performance alone is not enough. AI does not only have a performance and safety problem. It has a deployment problem.

Most AI systems fail not because they are technically insufficient, but because the organisations, regulations and infrastructures that must absorb them were not designed for them — and they were not designed for those constraints either.

A performant model is not a deployable system.

A deployable system is not necessarily economically sustainable.

A sustainable system must also be governable and controllable over time.

This is not an additional constraint. It is the precondition for any real industrialisation.

—

The RAISE Framework

A reference architecture for deploying AI in critical and regulated environments. Five interdependent pillars — applicable to any sectoral normative corpus.

Pillar	Title	Description
R	Regulatory Architecture	Regulatory frameworks integrated from the outset of system design
A	Accountability & Governance	Responsibility, auditability and institutional risk control
I	Interoperability Standards	Traceability, data sovereignty and system interoperability
S	Safety & Operational Validation	Operational validation, human oversight and reversibility
E	Explainability & Ethics	Algorithmic transparency and decision-making legitimacy

Healthcare & Life Sciences

EU AI Act · MDR/IVDR · HDS · SNDS · GDPR

Defence & Security

NIS2 · NATO standards · sovereign cybersecurity

Public Sector

RGS · GDPR · DORA · ANSSI frameworks

Critical Infrastructure

IEC 62443 · ISO 27001 · CER Directive

Finance & Insurance

Solvency II · DORA · Basel III

🔒 Access the RAISE Framework

—

Work Domains

Design & Architecture

AI systems architecture at scale

Organisational integration, observability, resilience, human oversight and reversibility.

Design & Architecture

Data engineering as infrastructure

Quality, traceability, interoperability and long-term stability.

Design & Architecture

System lifecycle & degradation

Operational monitoring, model drift, revalidation and full lifecycle governance.

Regulation & Governance

Compliance by design

EU AI Act, GDPR, MDR/IVDR and sectoral frameworks embedded in system architecture from day one.

Regulation & Governance

Algorithmic governance

Accountability, auditability, risk control and decision-making legitimacy.

Regulation & Governance

Digital sovereignty & geopolitics

Trusted cloud, data localisation, sovereign constraints in critical systems.

Scale & Viability

Performance evaluation & measurement

Defining, validating and standardising performance in real operational contexts.

Scale & Viability

Economics of industrial AI

Deployment economics in regulated environments, cost structures, clinical and institutional ROI.

Scale & Viability

Human & organisational transformation

Change management, executive acculturation, capability building and institutional resilience.

Explore work domains

—

Reference Projects

Research project · Healthcare & Life Sciences

PREDICARE

A territorial predictive medicine programme built on an architecture of individual medical digital twins. Derived from the Programme Sentinelle IA, PREDICARE structures the clinical scale-up of a predictive approach in primary care medicine.

Target budget€3.5M

Pilot phase~100 patients

FundingBPI France pre-validation

DeploymentGHT pre-validation · Regional funding committed

View project

Industrial project · Healthcare & Life Sciences

TweenMe

The first universal digital twin generator in Do It Yourself mode — designed for the domain expert who holds a research question and a data asset, with no prerequisite in machine learning. Delivers a reliable digital twin in a matter of weeks, with quality and lifecycle fully controlled end-to-end.

ByQualees

StatusUnder development

CertificationPlanned — not yet initiated

View project

View all projects

—

Publications

Doctrinal work, analyses and reference frameworks on the industrialisation of AI in critical systems.

Research Papers & Reports

🔒

Research paper

"Medical abandonment as a systemic property: structural diagnosis and predictive prevention infrastructure"

PREDICARE Memoir v3 — 2025 · 110 pages

🔒 Registration required Preview + access

Articles & Doctrinal Positions

○

Article — Doctrinal position

[Article title — to be specified]

[Abstract — 2 lines max] · Open access · PDF

○ Open access Read

Frameworks & Deployment References

🔒

Reference framework

RAISE Framework — Reference Document · Deployment architecture for AI in critical and regulated environments

Registration required

🔒 Registration required Preview + access

View all publications

—

Explorations

Doctrine does not emerge in a vacuum. It grows from a broader intellectual culture — philosophy of science, history of technology, economics of complex systems, epistemology of measurement.

These are the readings, reasonings and detours that feed the thinking without necessarily appearing in it directly. Open access. No registration required.

Explore

—

Jérôme Vetillard

Scientific & engineering background

Engineer — AgroParisTech

Biotechnology · Industrial bioreactors

PhD in Biotechnological Engineering — AgroParisTech

13th ENS Ulm national competitive entrance examination

Specialised medical training

Haematology-Oncology · Radiopathology & Radiation protection

Clinical oncology · Cell therapy (stem cells)

Genetic engineering · Nuclear Medicine

Treatment of acute radiation syndrome (haematopoiesis, GI barrier, inflammatory storm)

Management & strategy

Global Executive MBA — IE Business School / Brown University

Chief Product Officer Program — MIT Sloan (ongoing)

Career

20 years · Microsoft

Global Solution Architect (Business Productivity)
Principal Enterprise Architect (CTO)
Senior Digital Advisor — Regulated Industries (CDO)
Senior Industry Advisor — Healthcare & Life Sciences (CPO)

Digital transformation programmes up to €148M in licences and €28M in services
(manufacturing 4.0, government cloud, defence cloud…)

VP R&D & Chief Product Officer · Qualees

Founder · Twingital Institute

Positioning

At the intersection of complex systems engineering, specialised clinical medicine and product strategy in regulated environments.

Learn more

Twingital Institute

Independent research space. Development of doctrine, conceptual frameworks and deployment reference architectures.

Structural thinking.

Implementing frameworks in real operational contexts. Strategic advisory, deployment architecture, executive support in regulated industries.

Industrial execution.

Discuss an engagement

Industrial AI in critical, highly regulated environments.

Doctrine

The RAISE Framework

Work Domains

Reference Projects

PREDICARE

TweenMe

Publications

Explorations

Jérôme Vetillard

Your AI system performs well.Is it actually deployable?

Your AI system performs well.
Is it actually deployable?