Applied research in cyber & AI security
Work focused on architecture, governance, assurance and control of AI systems and autonomous agents, with strong emphasis on traceability and human authority.
Structural security problems.
The public presentation intentionally stays at architecture level: proprietary mechanisms, internal parameters and patent-oriented details are not published.
Cyber + AI + Agentic
Three-layer enterprise architecture: conventional cybersecurity, AI-system security, then autonomous-agent security and action capabilities.
Context, authority and action
Extending Zero Trust beyond human identity to context, data, tools, memory, delegated authority and action execution.
AI, agents & human approval
Decision traceability, authority levels, action gates, revocation, human control and evidence governance.
AI error measurement
Qualifying, measuring and tracing errors to support evaluation, decision-making and continuous improvement of AI systems.
Decision support & evidence governance
Connecting assets, signals, risks, decisions, controls and evidence to strengthen security steering.
Chain of trust
Revocation, trust reconstruction, monitoring and continuity when cyber, AI or agentic components become untrusted.
Research framed by governance and security.
The work is structured to remain usable in enterprise architecture and CISO governance contexts.
Assets, actors, threats, exposure, criticality and dependencies.
Architecture, accountability, decision points, isolation, observability and human action.
Validation criteria, logs, metrics, tests, decisions and traceability.