Marius Dima
What I Do
I build AI systems that explain themselves. For critical infrastructure where you need to know why a decision was made, not just what it decided.
20+ years in industrial automation, manufacturing, and AI. 3 published patents. Speaking at MIT, AI Summit London, How to Web.
My focus: Neural-symbolic architectures that merge deep learning with logical reasoning. EU AI Act compliance frameworks. Pattern recognition with full transparency.
Core Projects
Qriton Platform
qriton.comNeural-symbolic AI with EU AI Act compliance. Explainable decision-making for high-risk systems.
@qriton/hopfield-anomaly
npmjs.com/package/@qriton/hopfield-anomalyProduction-ready anomaly detection using energy-based models. Unsupervised pattern recognition with explainable outputs.
iGov.ro
igov.roOpen governance platform for Romanian democracy. Transparency tools for citizen participation.
AIRadio.Host
airadio.hostAlgorithmic broadcasting research. Content automation with full transparency.
Expertise
- Neural-Symbolic AI: Combining neural networks with symbolic reasoning for explainable systems
- EU AI Act Compliance: Frameworks for high-risk systems meeting regulatory standards
- Energy-Based Models: Hopfield networks for anomaly detection in critical infrastructure
- Industrial AI: Manufacturing automation, SCADA systems, predictive maintenance
- Stack: Node.js, Python, Vue.js, TypeScript, SQL/NoSQL, Docker, Cloud
Recent Talks
Contact
Available for consulting on neural-symbolic AI, explainable systems, and EU AI Act compliance.