Core contributor to MANGO (manycore HPC systems), RECIPE (exascale heterogeneous systems), and continued HPC work in INCISIVE.
CENTRO REGIONALE INFORMATION E COMMUNICATION TECHNOLOGY SCARL
Italian research centre specializing in HPC architectures, applied AI for healthcare imaging, and secure data processing for space and cloud systems.
Their core work
CERICT is a southern Italian research centre specializing in high-performance computing architectures, data privacy compliance, and applied AI for critical sectors. Their technical work spans designing manycore HPC systems for demanding workloads, building privacy-aware linked data platforms, and developing AI-based diagnostic tools for healthcare imaging. They bring strong computing infrastructure and data engineering expertise to EU consortia tackling problems from space system security to cancer detection.
What they specialise in
INCISIVE project applies XAI, deep learning, neural networks, and federated learning to breast, colorectal, and lung cancer imaging diagnostics.
SPECIAL focused on scalable privacy-aware linked data architecture; MUSA addressed multi-cloud security — both centered on data governance.
MUSA targeted multi-cloud secure applications; 7SHIELD addressed safety and security of space ground segments and satellite data.
7SHIELD project focused on protecting space system ground segments and Copernicus satellite data assets.
How they've shifted over time
CERICT started firmly in the HPC hardware and architecture space (2015–2019), working on manycore processors, real-time computing, power-performance optimization, and exascale system management through MANGO and RECIPE. From 2020 onward, they pivoted toward applying their computing muscle to domain-specific challenges: AI-driven cancer diagnostics with federated learning (INCISIVE) and space infrastructure security (7SHIELD). The shift signals a move from building computing platforms to deploying them for high-impact applications in health and security.
CERICT is transitioning from pure HPC infrastructure research toward AI applications in healthcare and secure data processing, making them increasingly relevant for projects that need heavy computation applied to real-world problems.
How they like to work
CERICT operates exclusively as a consortium participant — they have never coordinated an H2020 project, which suggests they position themselves as a technical contributor rather than a project leader. With 77 unique partners across 18 countries in just 6 projects, they work in large, diverse consortia and are clearly comfortable in multi-national, multi-disciplinary teams. This broad network without repeat coordination indicates they are sought after for specific technical capabilities rather than driving project agendas.
Despite only 6 projects, CERICT has collaborated with 77 distinct partners across 18 countries, reflecting their participation in large consortia spanning most of the EU. Their network is broadly European with no obvious geographic concentration.
What sets them apart
CERICT sits at the intersection of HPC infrastructure and applied AI — a relatively rare combination among Italian research centres. While many organizations do either computing research or domain applications, CERICT can bridge both: they understand hardware-level optimization (manycore, exascale) and can apply that knowledge to demanding AI workloads like medical image analysis. For consortium builders, they offer a partner that brings both the computational backbone and the expertise to deploy it in sensitive domains like health and security.
Highlights from their portfolio
- MANGOTheir largest-funded project (EUR 571K), focused on next-generation manycore HPC architectures — represents their core technical identity.
- INCISIVETheir most recent and thematically ambitious project, combining AI, federated learning, and XAI for multi-cancer diagnostics — signals their strategic direction.
- 7SHIELDUnusual cross-domain project applying security expertise to space ground segments and Copernicus satellite data, showing versatility beyond pure computing.