Core competency threaded across all four H2020 projects — their company name itself references Bayesian probability, and they contribute AI/ML components to diverse application domains.
PROBAYES SAS
Grenoble-based SME specializing in probabilistic AI and Bayesian inference for health, manufacturing, and emerging quantum/optical computing applications.
Their core work
Probayes is a French SME specializing in probabilistic artificial intelligence and Bayesian inference, based in the Grenoble technology cluster. They develop AI solutions that handle uncertainty and complex decision-making across diverse application domains — from assistive robotics for elderly care (GrowMeUp) to industrial process optimization (MONSOON) and optical computing architectures (COPAC). Their core value lies in applying advanced probabilistic modeling and machine learning to real-world systems where data is noisy, incomplete, or high-dimensional.
What they specialise in
MONSOON project focused on model-based control frameworks for site-wide optimization of data-intensive industrial processes.
GrowMeUp project applied AI to growing robotic capabilities for elderly assistance and independent living.
Participated in COPAC (coherent optical parallel computing) and contributed to GreQuE (Grenoble Quantum Engineering Doctoral Programme), signaling a move toward next-generation computing paradigms.
How they've shifted over time
Probayes began its H2020 participation with applied AI projects in health robotics (GrowMeUp, 2015) and manufacturing optimization (MONSOON, 2016), both grounded in classical probabilistic methods. By 2017, their focus shifted toward quantum engineering, optical computing, and nanotechnologies — as evidenced by their involvement in GreQuE and COPAC. This trajectory suggests a deliberate move from applying existing AI methods to industrial problems toward exploring next-generation computing architectures where probabilistic reasoning meets quantum and photonic systems.
Probayes appears to be positioning itself at the intersection of probabilistic AI and emerging computing paradigms (quantum, optical), which could make them a valuable partner for projects exploring AI algorithms on non-classical hardware.
How they like to work
Probayes consistently joins projects as a participant or third-party contributor rather than leading consortia — none of their four projects were coordinated by them. With 44 unique partners across 13 countries from just four projects, they operate in relatively large consortia and appear comfortable working with diverse, international teams. This profile suggests a specialist contributor that brings focused technical expertise to larger collaborative efforts rather than driving project-level strategy.
Despite only four projects, Probayes has built a broad network of 44 unique consortium partners spanning 13 countries, indicating they join sizable international consortia. Their Grenoble base places them in one of France's strongest tech and research ecosystems, likely facilitating connections to CEA, CNRS, and local university labs.
What sets them apart
Probayes occupies a rare niche as an SME focused specifically on probabilistic and Bayesian AI — not generic machine learning, but principled uncertainty modeling. Their Grenoble location embeds them in a world-class ecosystem for quantum technologies and advanced computing. For consortium builders, they offer a specialized AI partner that can bridge the gap between theoretical probabilistic methods and practical applications in health, manufacturing, or emerging computing domains.
Highlights from their portfolio
- COPACLargest EC contribution (EUR 367,500) exploring coherent optical parallel computing — an unusual intersection of AI and photonic hardware.
- GreQuEGrenoble Quantum Engineering Doctoral Programme connecting Probayes to the regional quantum technology ecosystem as a third-party industry partner.
- MONSOONModel-based optimization of data-intensive industrial processes — demonstrates direct applicability of their probabilistic AI to manufacturing use cases.