Résumé

Nicolas Pagès

Data & AI transformation consultant · 20 years of industrial leadership

Twenty years between industry and software: chip testing, medical-device industrialisation, factory 4.0, then Machine Learning and AI agents. Today I design and run applications that work for SMEs, at NPIA Lab.

Ajaccio, Corsica, remote anywhere

Projects

What I build and run right now, read live from NPIA Lab's own management tool.

NPIA-OS : interactive mockup · fictional data

NPIA-OS

Laboratoire de R&D interne de NPIA Lab, et poste de pilotage de sa propre activité : acquisition, devis et signature, projets et temps, finance, trésorerie et business plan, et portail client (suivi du projet étape par étape, signature des devis, factures, documents, support). Assistant IA intégré (Aidy) qui propose et attend l'accord humain avant d'agir.

Dec 2025 -- Mars 2026Open the mockup
PythonNext.jsPostgreSQLAgents IAAPI LLM
InvoiceRisk AI : interactive mockup · fictional data

InvoiceRisk AI

SaaS de prédiction des retards de paiement : score de risque par facture (heuristique puis modèle ML par organisation), relances priorisées en trois tons, mesure de l'impact sur le DSO et le recouvrement.

MLPythonAPI RESTNext.jsStripe
Coulisses : interactive mockup · fictional data

Coulisses

Outil de coordination d'une salle de spectacle associative : agenda, fiche par date avec attendus calculés et feuille de route, suivi des manques, classement automatique des documents, planning partagé.

Next.jsPostgreSQLClassement automatiqueCalendrier .ics
Rosaventia Retail : interactive mockup · fictional data

Rosaventia Retail

ERP de commerce indépendant connecté à la caisse : stock par mouvements, réception lue par IA, achats, valorisation, clôture ; complété d'un site e-commerce, d'une app vendeur PWA hors-ligne et d'une app client fidélité.

Next.jsPostgreSQLPWAConnecteur caisseLecture de documents
Vigie : interactive mockup · fictional data

Vigie

Client confidentiel

Système de veille assistée par IA : collecte automatique de sources publiques, tri, résumés avec citations, rédaction et publication d'un bulletin sur blog privé. Conçu comme veille réglementaire pour un client, généralisé à tout domaine.

CrawlingLLMPythonNext.jsWorkflow éditorial
Analivia : interactive mockup · fictional data

Analivia

SaaS de recherche qualitative : transcription, verbatim, structuration thématique et remplissage automatique des grilles d'entretien avec citations vérifiables. Pseudonymisation réversible et crédits à la minute.

NLPTranscriptionPythonNext.jsRGPD

Plateforme interne d'un distributeur électronique

Client confidentiel

Plateforme web métier : analyse de références et BOM depuis des sources priorisées, catalogue, CRM, tableaux de bord commerciaux (CA, objectifs, book-to-bill), module terrain avec comptes rendus dictés et mesure d'impact.

PythonNext.jsPostgreSQLScoringSpeech-to-text
Studio : interactive mockup · fictional data

Studio

Client confidentiel

Plateforme de création et de planification de contenus réseaux sociaux à partir d'un catalogue produit : enrichissement automatique, propositions visuelles et textuelles par IA, revue humaine, calendrier éditorial.

LLMGénération d'imagesPythonNext.jsPipelines

Experience

Twenty years, six companies, one thread: taking a technical idea into production.

  1. Feb 2026 - today

    CEO, Data & AI transformation consultant

    NPIA Lab · Ajaccio

    Consulting and engineering practice: strategic framing, use-case prioritisation, design and production release of the selected solutions, team onboarding. Working with owners of SMEs from 10 to 250 employees.

    • Electronic components distributor: sales-process diagnosis, decision layer on top of the ERP, part-analysis and equivalence engine, field module for sales reps. Phase 1 signed and being deployed.
    • Beauty products distributor: product-repository overhaul, AI-assisted content enrichment and multi-channel publishing pipeline. Platform delivered, in continuous operation.
    • Products: Analivia (qualitative interview analysis), NPIA-OS (business steering with an AI assistant), and an execution foundation for AI processing: event log, cost charged per step, recovery after incident, human approval of committing actions.
    • Ongoing R&D projects: two ERPs published by NPIA Lab, Rosaventia Retail (independent retail, wired to the point of sale) and Rosaventia Consulting (multi-tenant steering for consulting firms, freelancers and liberal professions, derived from NPIA-OS).
    Machine LearningAI agentsSaaS
  2. Sep 2021 - Nov 2024

    Chief Operating Officer, executive committee member

    Telaqua · Marseille

    Agritech: connected irrigation control. Technical and industrial leadership for connected equipment deployed on farms.

    • Design of the factory 4.0 and industrialisation of connected equipment.
    • Automation of functional tests and accessory configuration.
    • Statistical analysis of production test data: detection of manufacturing drifts, identification of the least reliable components.
    • Management of supply chain, logistics, inventory and production.
    • Management of industrial R&D projects and suppliers.
    Industrial R&DFactory 4.0Data analytics
  3. Sep 2020 - Aug 2021

    Deputy COO

    Nexialist · La Ciotat

    Medical-device consulting firm (regulatory affairs and quality).

    • Managed a team of twenty expert consultants.
    • Overhauled business steering: automated decision reporting (Power BI) tied to the ERP.
    • Structured operational processes.
    ManagementMedical devicesAutomation
  4. Jan 2016 - May 2022

    Founder and principal consultant

    Ariga Solutions · Gardanne

    Consulting practice in industrialisation and process automation, from deep-tech startups to subsidiaries of international groups and public research labs. Part-time from September 2020.

    • "Industry of the future" diagnostics for a metallurgy trade body: maturity assessment against a reference framework, prioritisation crossing stakes and accessibility, costed roadmap, debrief to the owner.
    • Design of automated production lines with robotics and computer vision for deep-tech startups (agrivoltaics, service robotics, scientific cameras).
    • Framing of data and AI use cases in industry: predictive maintenance on sensor data, RFID/NFC traceability tied to the ERP, digitisation of purchasing and production flows.
    • Medical-device manufacturer, subsidiary of an industrial group: coaching of industrial teams over nearly three years.
    • Industrial materials group: six months of R&D and industrialisation support on the design of a construction robot.
    • Public astrophysics laboratory: one year supporting the formalisation and documentation of a patented sensor-curving process ahead of its transfer to industry; the technology led to a spin-off and a commercial curved sensor.
    • RFID identification manufacturer: design, build and deployment of a tag reading and configuration robot on a four-axis arm.
    • Outdoor equipment manufacturer: design and build of an automated machine producing thermo-folded enclosures.
    • Hyperbaric exposure safety engineering: production line and CE marking.
    IndustrialisationRoboticsComputer vision
  5. Jun 2008 - Dec 2015

    Head of Industrialisation, executive committee member

    SuperSonic Imagine · Aix-en-Provence

    Ultrasound imaging medical devices.

    • Industrialisation of equipment, from R&D to series production.
    • Statistical use of production data to steer quality and yield.
    • Regulatory certifications and transfer of production abroad.
    Medical devicesCertificationsProduction
  6. Nov 2004 - Jun 2008

    Test and Product Engineer

    Texas Instruments · Nice

    Semiconductors.

    • Development of component test protocols and design of experiments.
    • Statistical analysis of production data (yield, drifts, correlations) for industrial decisions.
    SemiconductorsTestStatistics

Skills

What I do, grouped by family.

Data Science & IA

  • Machine Learning
  • Deep Learning
  • NLP
  • Systèmes de recommandation
  • Séries temporelles
  • MLOps

Développement & API

  • Python
  • API REST
  • Next.js
  • PostgreSQL
  • Architecture multi-agents
  • Intégration LLM

Gestion de Projet R&D

  • Pilotage de projets techniques
  • Spécifications fonctionnelles
  • Budgets R&D

Automatisation & Systèmes

  • Dashboards custom Python/Next.js
  • ERP
  • Robotique industrielle
  • Usine 4.0
  • Systèmes embarqués

Business Intelligence

  • Power BI
  • Looker Studio
  • Tableau
  • Reporting automatisé

Education

The foundation, and the 2026 update.

2025 - 2026

AI Engineer

OpenClassrooms

Professional certification RNCP level 7 (master's level), 804 supervised hours, twelve projects defended before a jury: machine learning, deep learning for text and images, MLOps and production deployment of an inference engine, distributed processing on the cloud, framing of an AI project presented to an executive committee.

See the syllabus
2001 - 2004

Engineering degree

Centrale Méditerranée (ENSPM), Marseille

Fundamental physics, optics, signal processing, computer science, microelectronics.

1999 - 2001

Two-year degree in Electrical Engineering and Industrial IT

IUT Nice Côte d'Azur

Electronics, automation, industrial computing.