We are proud to share that Mattheos Fikardos has successfully defended his thesis: “Trustworthy AI Operations with Conversational Assistance and Agentic Workflows” — a landmark piece of work that will shape how we think about deploying AI responsibly at scale.

The hesis establishes the discipline of Trustworthy AI Operations, providing a roadmap for autonomous AI systems that are not just powerful, but reliable, transparent, and aligned with human values.

At the heart of his contribution is a novel AI operational maturity taxonomy — five evolving levels that chart the journey from manual implementations to self-governing multi-agent systems:
🔧 Handcrafted – manual, disjoint implementations with minimal automation
🔗 Integrated – unified lifecycle management eliminating data and model silos
🤖 Augmented – LLMs and agentic AI lowering the cognitive barrier to complex AI system management
✅ Trusted – continuous verification of fairness, robustness, explainability, and other trustworthiness requirements
🌐 Autonomous – self-governing multi-agent systems operating within established trust boundaries.

To support this vision, Mattheos developed a reference architecture that unifies AI development stages and reduces technical debt, and defined the Trustworthiness Optimisation Process (TOP), a method that treats sociotechnical constraints as live operational objectives rather than one-off compliance checkboxes.

As a technical validation of these concepts, he built Trustproofer, a neurosymbolic agentic framework blending symbolic reasoning with neural adaptability. Trustproofer serves as a practical testbed, demonstrating that high-level autonomy can coexist with rigorous human oversight and sociotechnical safeguards.

With 8 peer-reviewed publications (4 journals and 4 conferences) and important contributions to research projects like THEMIS 5.0, Wasabi EU Project, and COALA Your Factory Assistant, Mattheos has already made a significant mark on the research community.

He has also presented his work at the Trustworthy AI Summit of the European Trustworthy AI Association, the TAILOR Network of Excellence Centres on Trustworthy AI, and the etami taskforce of the BDVA – Big Data Value Association.

Congratulations Matheos on this milestone!

Here’s to shaping the future of trustworthy AI.