UAIF™
Unified AI Incident Framework for structured incident classification and taxonomy.
AI securityApplied researchStandards
Mukesh Kumar Singh
Turning AI governance into operational assurance.
AI Security Researcher, Framework Architect, and Founder of ODA3 Institute.
I work at the intersection of AI security, incident response, evidence, standards development, and assurance—translating high-level requirements into systems that can be implemented and assessed.
21+Years of research
2005First publication
3ODA3 framework suite
4Defensive publications
02 / Profile
Research shaped by operational reality.
For more than two decades, my work has moved across enterprise security, government environments, digital forensics, threat intelligence, and applied AI security research.
That range informs a practitioner-first approach: define the problem precisely, identify the evidence that can support a conclusion, and design controls that survive contact with real systems.
Today, my focus is the operational layer between AI governance standards and observable system behaviour—particularly incident classification, response coordination, security assurance, and assessment-ready evidence.
03 / Current work
ODA3 Institute
Building the operational assurance infrastructure for AI.
ODA3 Institute is an applied research and advisory firm connecting AI governance requirements with controls, incident practices, evidence, assessment, and certification infrastructure.
AI-IRF™
AI Incident Response Framework for coordinated, evidence-aware response activity.
GAISSF™
Global AI Security & Safety Framework for governance, controls, and assurance.
Clear attribution
UAIF™, AI-IRF™, GAISSF™, and associated institutional publications are assets of ODA3 Institute. This personal site presents my role and research context; authoritative framework materials are published by the Institute.
Visit ODA3 Institute ↗04 / Research archive
Selected research and practitioner work.
The archive records work across AI security, threat intelligence, detection, digital forensics, and foundational system security. Publication labels describe the original works; they do not imply current product status or third-party validation.
05 / Working principles
Evidence before assertion.
Four disciplines guide the work.
- 01
Define the claim
Separate observation, inference, and recommendation so readers can see what the evidence actually supports.
- 02
Design for operations
Translate principles into accountable actions, artefacts, decision points, and measurable outcomes.
- 03
Preserve traceability
Maintain the connection between requirements, controls, implementation, evidence, and assessment.
- 04
Record what did not happen
Document notable absences and data limits rather than filling uncertainty with inflated conclusions.
06 / Collaborate
Research · Standards · Enterprise evidence
For work that benefits from technical depth and candid evidence.
I welcome focused conversations with researchers, standards participants, security leaders, and organizations working on AI-security implementation and assurance.
Start a conversation ↗