St. Pölten, Austria
The CERTAIN project marked a significant milestone this week as partners gathered at the University of Applied Sciences St. Pölten (SPU) for an intensive technical workshop, followed by a featured presentation at the SAINT Conference. The day’s events focused on bridging the gap between complex AI regulation and technical implementation.
Accelerating technical integration: The CERTAIN Workshop
Partners from SPU, Digital for Planet (D4P), and Empower (EMPW) held a productive day-long session aimed at synchronizing the project’s technical core. A primary focus was the roadmap for Pilot 3 (Energy), ensuring a full end-to-end integration of the CERTAIN tools by September 2026.
Key technical advancements discussed include:
- Semantic MLOps Engine: D4P presented the current architecture of the engine and the Postgres database designed to capture vital metadata for compliance checks.
- Ontology & Knowledge Graphs: SPU showcased the AIDOC-AP ontology. The teams aligned on discussing the current mapping between the Engine database and the ontology, as well as the transformation of the raw data into a virtual knowledge graph.
- Synthetic Data Generation: Partners defined a joint strategy. By combining household and grid-level generators (D4P) with OBIS-standard schemas (SPU/Empower), the project will create robust datasets to scale energy-community scenarios.
The workshop concluded with a visit to SPU’s new AI Real-World Laboratory, providing partners with a first-hand look at the infrastructure where the outcomes of CERTAIN are expected to be used.
"Compliance built into the pipeline": CERTAIN at SAINT 2026
Following the workshop, the momentum moved to the SAINT Conference, where Dimitrios Christodoulou (D4P) presented a compelling vision for the future of trustworthy AI: “Compliance Built into the ML Pipeline.”
The presentation addressed a critical bottleneck in the European AI market: The reliance on manual, static audits, which stifle innovation and increase costs. Dimitrios detailed how CERTAIN is transforming dense legal texts (such as the EU AI Act and GDPR) into actionable technical constraints.
The CERTAIN blueprint for trust:
- Semantic MLOps Engine: Automated logging of the full AI lifecycle.
- Ontology: Bridging the gap between technical artifacts and legal language.
- RegOps Tool: Continuous, automated assessment of bias and data quality.
By automating these processes, CERTAIN is ensuring compliance and lowering the barrier to entry for European SMEs to develop scalable, sustainable, and fundamentally trustworthy AI solutions.