Technology

A research thesis in four layers.

Perception, sensing, reasoning, evidence. Four layers engineered in Turin for the specifics of Italian DOP, IGP and Bio and French AOP, AOC and Label Rouge supply chains. Each layer carries its current maturity so nothing reads as shipped when it isn't.

  1. 01 · Computer vision · Spectral imaging
    In prototype

    Perception

    Vision models trained on cellar, mill and packing-line footage to detect substitution, mislabeling and adulteration in raw and packaged product. Tuned for the visual grammar of Italian and French terroir.

  2. 02 · IoT · Environmental telemetry
    In prototype

    Sensing

    Reference hardware engineered in Turin to capture temperature, humidity, geolocation and chain-of-custody events at every hop of the supply chain, from field to shelf.

  3. 03 · Federated ML · Privacy-preserving
    In research

    Reasoning

    Federated learning across Italian and French consortia to train stronger models without ever moving raw data. Sovereign, EU-hosted, GDPR-native by construction.

  4. 04 · Signed provenance · APIs
    In design

    Evidence

    Every verdict is packaged as a structured provenance dossier: images, sensor logs, model version, timestamps and operator identity, digitally signed so it is tamper-evident. Consortia can attach it to certification-body audit files or submit it to a court-appointed expert. It does not replace a ruling, but it supplies reproducible technical evidence.

Principles

Sovereign

European infrastructure, European law. Producer data never leaves the consortium or the country of origin without explicit consent.

Auditable

Every automated verdict is reproducible and reviewable by domain experts.

Interoperable

REST and event APIs plug into ERP, MES and consortium platforms already in use.

Want to shape the architecture?

We share architecture briefings and co-design proposals with consortia, researchers and public bodies interested in joining the first pilot cohort.

Talk to the founding team