AI Agents
How can specialized agents coordinate dependent tasks across enterprise applications?
Research
Enterprise AI &
Agent Security.
How can enterprise AI agents act across applications while preserving authorization boundaries, observability, and operational control?
This question guides my research interests in enterprise AI and connects them with my published work and practical architecture writing.
Questions I’m exploring
Research interests
My interests connect autonomous agents with the security and operational requirements of enterprise systems.
Agent Authorization
How can an agent preserve a user’s identity, access permissions, and delegated authority throughout a workflow?
Agent-to-Tool Security
How should tool access, API calls, and execution permissions be scoped and checked before an agent acts?
Enterprise AI Security
How can identity, data protection, and trust boundaries remain clear across models, integrations, and enterprise applications?
AI Observability
What evidence helps explain an agent’s decisions, tool calls, and outcomes during operation and investigation?
Secure-by-Design AI
How can security assumptions, approval points, and failure handling be considered early in an agent’s design?
Publications
Published work with inventor attribution, publication details, and original source records.
Architecture in practice
Related technical writing
Public technical articles and reference implementations that explore aspects of this research theme, with links to their original publications.
Conversational Security in Oracle Fusion: An AI Agent for Least-Privilege Discovery
Presents a reference implementation that translates natural-language security questions into read-only Fusion privilege and role analysis through controlled integrations.
Oracle A-TeamFusion AI Agents: Secure by Design Best-Practice Considerations
Outlines architectural considerations for Fusion agents, connecting business risk, identity context, instruction boundaries, segregation of duties, and available security controls.
Oracle Cloud Infrastructure BlogFusion AI Agents: Secure by Enforcement Best-Practice Considerations
Examines runtime security patterns for Fusion agents, covering least privilege, constrained tool use, approval boundaries, and controls beyond model behavior.
Oracle Cloud Infrastructure BlogFusion AI Agents: Secure by Operations Best-Practice Considerations
Discusses operational architecture considerations for Fusion agents, including observability, sensitive data handling, adversarial testing, change governance, and containment controls.
Oracle Cloud Infrastructure BlogReference implementations and architectural considerations are not statements of product availability or support. Each original publication provides its full context and attribution.
Research profiles
Find my work
Public research profiles and researcher identification.