Publications

Research memoir · PREDICARE · Forthcoming

"From medical abandonment to predictive medicine — structural diagnosis and territorial prevention infrastructure"

110-page research memoir documenting the complete intellectual and operational arc of the PREDICARE programme: epidemiological and economic foundations of medical abandonment as a systemic property, architecture of the territorial predictive prevention framework, clinical validation methodology, and governance model. Publication date to be announced.

Conference talk · AI in healthcare · VivaTech 2019 · 🇫🇷 FR & 🇬🇧 EN

« AI in Health: Past, Present, Perspectives »

Conference talk at VivaTech 2019 (Microsoft Health & Life Sciences). A structured retrospective on the trajectory of AI in medicine, the state of the art in 2019, and the conditions required for industrial-scale deployment. The earliest public formulation of the problem space that TweenMe and PREDICARE resolve today. Available in both English and French.

Reference framework · RAISE Framework

"RAISE — Reference architecture for AI deployment in critical and regulated environments"

Complete reference document for the RAISE Framework. Detailed architecture of the five pillars (Regulatory, Accountability, Interoperability, Safety, Explainability), sector-by-sector evaluation grids, interdependency matrices and operational implementation protocols.

Article · Healthcare system · 2018 · Series Part 1/3 · 🇫🇷 FR

« Le système de santé français : rappel d'architecture »

Mapping of the French healthcare ecosystem — institutions, agencies, care providers, payers, industrial actors. ONDAM, fee-for-service (T2A), PMSI coding. Structural growth of healthcare expenditure beyond GDP growth (11% of GDP). Analytical foundation for the PREDICARE programme.

Article · Health economics · 2018 · Series Part 2/3 · 🇫🇷 FR

« L'économie de la santé : principes et modèles de financement »

Health as a "superior good" (Kenneth Arrow, 1963). Comparative analysis of Bismarck vs Beveridge models and three welfare state typologies (Esping-Andersen). Aside on the US model: 18% of GDP for below-average outcomes, private management costs 5–6× higher than federal programmes.

Article · Health economics · Demographics · 2018 · Series Part 3/3 · 🇫🇷 FR

« Les causes inflationnistes des coûts du système de santé »

Technological progressivism and refusal of death as cost drivers. SOVALDI™ (Gilead, $75,000/treatment) as a case study in pharmaceutical pricing logic. Socio-demographic pressures quantified: Alzheimer (40→135M cases by 2050), cancer (+70% in 20 years), metabolic syndrome pandemic. Foundational document for PREDICARE.

Article · Healthcare systems & Digital innovation · 2015

« Sustainable healthcare services through digital innovation »

The structural unsustainability of reactive, acute-centred healthcare systems — and the case for applying cloud architectural patterns to health service delivery. Written in June 2015, this is the intellectual precursor to the PREDICARE programme and TweenMe. The diagnosis it establishes remains fully valid today.

Article · Healthcare systems & Digital innovation · 2016 · Part 2/2

« Sustainable healthcare services through digital innovation — Part 2 »

The architectural proposal: applying NIST cloud patterns to healthcare — standardised managed services, IoT/AI self-service, elastic supply chain, ubiquitous access. The 2016 blueprint that TweenMe and Programme Sentinelle IA operationalise nine years later.

Article · AI & Philosophy of science · April 2023 · 🇫🇷 FR

« L'intelligence artificielle : Est-ce aussi « con » qu'une planche de Galton ? »

Structural analogy between the Galton board (binomial law, conditional probabilities, Pascal's triangle) and neural networks (perceptron, hidden layers, synaptic weights). Transformer architecture (encoder/decoder, multi-head attention, Q/K/V matrices). Emergence of complexity through parameter massification. Limitations of layered architecture ("lasagna") versus the brain's fractal architecture. Implications for ethical AI and explainability.

Article · AI Architecture & Healthcare · 2025

« Clinically-Informed Neural Networks (CINNs): when medical literature becomes a learning constraint »

Transposing the PINNs principle to the clinical domain. Published biomedical literature — with its level of evidence — is integrated as a constraint in the loss function, not as training data. Formal distinction from classical regularisation. Original Qualees architecture, theoretical framework and epistemological positioning.

Academic article · Epistemology of AI · 2025

« Encoding, transduction and world models »

Biological encoding is transductive — direct causal coupling with the physical world via specialised sensory organs. LLM encoding is mediated by prior human decisions. Three irreducible levels of grounding. The polycentric memory graph as an architecture fundamentally distinct from statistical co-occurrence.

Article — Position paper · AI & Healthcare · 2025

« Bayesian: the word that buys regulatory credibility without earning it. »

Why most Bayesian implementations in health AI are frequentism dressed up as a probabilistic graph. Priors chosen for computational convenience, uncertainty not propagated, correlation/causality confusion. The structural HR argument: the architectural choice is partly a recruitment choice disguised as a technical choice.

Article — Doctrinal position · AI & Energy

« The supposed virtuality of AI facing the wall of reality: how $200 billion in GPUs reveal the physical constraints of artificial intelligence »

Industrial AI consumes growing amounts of energy that existing infrastructure cannot absorb indefinitely. Analysis of real physical constraints — datacentres, water, electrical grid — and their strategic implications for organisations deploying AI at scale.

Article — Doctrinal position · Digital Twins · 2025 · 🇬🇧 EN

"Do Buzzwords Dream of Clearer Substance? The Journey of Digital Twins"

Digital twins between buzzword and substance. Analysis of the conceptual journey of the digital twin — from industrial simulation to patient digital twin — and the conditions for architectural credibility in regulated environments.

Article — Doctrinal position · Agentic AI · 2025 · 🇬🇧 EN

"AI Never-Ending Story: Agentic Frameworks and the Tale of Self-Reinvention"

Agentic frameworks as the latest iteration of AI's never-ending story. Critical analysis of multi-agent architectures, algorithmic autonomy and the industry's perpetual self-reinvention narrative.

Article — Doctrinal position · Digital Twins · Healthcare · 2025 · 🇫🇷 FR

« The patient digital twin: anatomy of a concept »

Anatomy of the patient digital twin concept. Beyond the buzzword, what it actually covers — and the architectural, epistemological and regulatory conditions for it to have clinical meaning.

Article — Doctrinal position · Explainability · Healthcare AI · 2025 · 🇫🇷 FR

« Demanding AI explainability: a human vanity? »

Is the demand for explainability a legitimate regulatory imperative or a poorly framed anthropomorphic projection? Analysis of the tension between algorithmic transparency, clinical performance and regulatory constraints (MDR, EU AI Act).

Article — Doctrinal position · AI Architecture · Healthcare · 2025 · 🇫🇷 FR

« AI in healthcare is not just a technical problem — it is above all an architecture problem »

The failure of most healthcare AI projects is not technical — it is architectural. Analysis of structural deployment conditions: interoperability, data governance, operational sustainability.

Article — Doctrinal position · Cognitive load · Clinical AI · 2025 · 🇫🇷 FR

« Does AI reduce clinicians' cognitive load — or increase it? »

AI is supposed to help clinicians. But every tool added is another screen, another alert, another score to interpret. Analysis of the cognitive load induced by clinical decision support systems.

Article — Position paper · Clinical Trials & AI · 2025 · 🇬🇧 EN

"AI-Enabled Clinical Trials: The 2025 Evidence Engineering Framework"

A continuous evidence engineering framework combining adaptive clinical trials, synthetic controls, and traditional RCTs under unified governance. Enabling AI systems to evolve at software speed while maintaining regulatory-grade causal proof.

Article — Doctrinal position · AI & Bio-industries · 2024 · 🇫🇷 FR

« Impact of Artificial Intelligence (AI) on bio-industries »

Analysis of AI's impact on bio-industries — biotechnology, biopharma, agro-industry, bioenergy. Value chain transformation, drug discovery acceleration, and strategic implications for industrial players.

Article — Doctrinal position · Generative AI · Clinical trials · 2024 · 🇫🇷 FR

« Clinical Cohort Augmentation via Generative AI: An Innovative Approach to Bias Correction »

How generative AI can correct representativeness bias in clinical cohorts. Synthetic data augmentation, balancing under-represented sub-populations, and implications for trial validity.

Article — Doctrinal position · Healthcare AI · Health economics · 2025 · 🇬🇧 EN

"AI in Healthcare: Impressive Progress, Missing Proof"

No clinical AI should be scaled without independent health-economic evidence demonstrating its value for the common good — not for enriching a monopolistic hyperscaler. Impressive technical progress does not exempt from economic and clinical proof. That proof is systematically missing.

Article — Doctrinal position · Sovereign cloud

"Cloud strategy for French public administrations: comparative analysis using the GVI/PCI framework"

Analysis of strategic cloud options for French public organisations using the GVI/PCI analytical framework from the MIT CPO Program. Digital sovereignty, interoperability and regulatory constraints in government-scale cloud architecture decisions.

Scientific poster · ISPOR Glasgow 2025

"Synthetic population generation for a therapeutic trial in lung oncology — TweenMe Smart Data Fertilizer algorithm"

Presentation of TweenMe's Smart Data Fertilizer module applied to synthetic population generation in a lung cancer therapeutic trial. Resolution of the HDLSS (High Dimensionality Low Sample Size) problem through structurally informed fertilisation. Rated top 5% of posters by the ISPOR Glasgow 2025 international jury.

Conference · AFCRO · Clinical trials · 2025

AFCRO 2025 Conference

AI in clinical trials: methodological approaches, synthetic control arms, and deployment conditions within a demanding regulatory framework.

Conference · AFCRO · TweenMe · 2024

AFCRO 2024 — TweenMe Presentation

Presentation of the TweenMe digital twin generator. Data pipeline architecture, dimensional enrichment approach and first clinical validation results.

Conference · Haemato-oncology · AbbVie

Haemato-oncology presentation for AbbVie

AI and modelling in haemato-oncology. Convergence between clinical research and computational approaches in the treatment of haematological malignancies.

Teaching · EBI · Epidemiology · Clinical research

Epidemiology & Clinical Research Minor — EBI Course

91-slide course covering the fundamentals of epidemiology and clinical research: drug development, clinical trial methodology, biostatistics, regulation. Delivered at the École de Biologie Industrielle (EBI). Embedded viewer and gated download available.

Executive MBA · IE Business School / Brown University · 2015

Cape Town Project — Sanitation Services Access in Townships

Field study (Khayelitsha, Cape Town) on sanitation services access in informal settlements. Systemic interdependency analysis, proposed principles and experimentations. Embedded viewer and gated download available.

Innovation · Strategy · 2016 · 🇫🇷 FR

« Frugal innovation as a response to a VUCA world »

Analysis of the frugal innovation paradigm — doing more with less — in a Volatile, Uncertain, Complex and Ambiguous world. Case studies and implications for innovation strategies in large organisations.

Virology · COVID-19 · March 2020 · 🇬🇧 EN

« Reflections on some very nasty little things — Episode 1: What is a virus? »

Fundamentals of virology: viral structure, replication cycle, cell tropism. Written in March 2020 for a non-specialist audience, when understanding of SARS-CoV-2 was still being constructed.

Diagnostics · COVID-19 · March 2020 · 🇬🇧 EN

« Episode 2: Diagnosis kits & testing strategy »

RT-PCR kits, screening strategies, sensitivity/specificity, predictive value. Implications for public health policy during a pandemic.

Physiopathology · COVID-19 · April 2020 · 🇬🇧 EN & 🇫🇷 FR

« Episode 3: Physiopathology of COVID-19 »

Immune response, cytokine storm, acute respiratory distress syndrome, multi-organ damage. Available in English and French.

Geriatrics · Nursing homes · COVID-19 · April 2020 · 🇫🇷 FR

« Why mortality is so high in nursing homes and why palliative care is proposed »

Excess mortality in French nursing homes during the COVID-19 crisis. Geriatric risk factors, limitations of invasive care, ethics of palliative care in a pandemic context.

AI & Consciousness · Philosophy of mind · 2024 · 🇫🇷 FR

« Artificial intelligence and the question of "consciousness" — Chapter 1: The Glasgow Coma Scale »

Can the Glasgow Coma Scale be applied to AI? Before discussing the emergence of consciousness and singularity, what emergency medicine, forensic psychiatry and philosophy of mind actually say about consciousness — and what it implies for AI.

Quantum physics · Popular science · 2024 · 🇫🇷 FR

« A ladle of quantum physics, anyone? (ladle #1) »

Introduction to quantum mechanics for non-physicists. First instalment of a popular science series on the fundamental concepts of quantum physics.

Quantum physics · Popular science · 2024 · 🇫🇷 FR

« A ladle of quantum physics, anyone? (ladle #2) »

Continuation of the quantum mechanics exploration. Deepening the concepts introduced in the first ladle.

Quantum computing · 2025 · 🇫🇷 FR

« Majorana fermions: To qubit or not qubit? »

Majorana fermions — particles that are their own antiparticle — as candidates for topological qubits. Stakes for fault-tolerant quantum computing.

Quantum computing · Error correction · 2025 · 🇫🇷 FR

« Qubits under Control: The Great Challenge of Quantum Stability »

The central challenge of quantum computing: maintaining qubit coherence long enough to perform useful computations. Decoherence, error correction, tolerance thresholds.

Quantum physics · Reflection · 2024 · 🇫🇷 FR

« A few answers.. »

Follow-up to the quantum physics series. Clarifications, nuances and extensions.