Responsible AI Leadership

Building AI that deserves people's trust.

Rani Kumar Rajah is a Responsible AI strategist, founder of SecuraAI, and creator of SafeEval. Her work spans AI Trust & Safety, youth safety, agentic AI, evaluation, governance, education, and enterprise AI strategy.

Explore the work Open knowledge

Rani speaking at TiE Silicon Valley — TiECon Silicon Valley.

20+Years technology leadership

600+Professionals trained

5+Standards & research contributions

2Interactive OWASP comic adaptations

Trust is not a feature. It has to be designed, evaluated, and continuously earned.

My work focuses on turning emerging AI risks into practical product decisions. I bring together enterprise leadership, Responsible AI strategy, mental health AI safety, agentic AI research, visual education, and hands-on training to help organizations deploy AI systems that are capable, useful, and demonstrably safer for the people who depend on them.

Featured Work

A body of work across AI evaluation, youth safety, agentic AI, standards, and education.

Creator

SafeEval

A Responsible AI evaluation framework for youth mental health AI, focused on realistic multi-turn conversations, clinically informed personas, trajectory-based metrics, and safety validation before deployment.

Time-to-DetectionTrajectory AccuracyEscalation PrecisionTurn-Level Recall

SafeEval →

Research Lead

Project FERAL

Open research mapping agentic AI architectures to CSA MAESTRO, OWASP ASI Top 10, and MITRE ATLAS with real-world validation and hardening guidance.

Project FERAL →

Creator

Mental Health AI Safety Library

Realistic safety scenarios for therapy chatbots, AI companions, coaching assistants, and clinical AI applications.

Creator

Interactive OWASP Comic Adaptations

Visual learning experiences that make OWASP LLM Top 10 and OWASP Agentic Top 10 concepts accessible without sacrificing technical depth.

LLM Top 10 Comic →
Agentic Top 10 Comic →

Making Responsible AI Accessible

Research, tools, and learning assets that help teams understand, evaluate, and act.

OWASP LLM Top 10 Comic

Interactive visual adaptation for developers, students, security practitioners, and GRC audiences.

Explore → 02

OWASP Agentic Top 10 Comic

Comical adaptation of agentic AI risks including tools, autonomy, identity, memory, and trust.

Explore → 03

AI Safety for Kids

A 5Cs framework-based initiative introducing children and educators to AI safety.

Explore → 04

AI Security Challenge Arena

A CTF-style environment for realistic hands-on AI security challenges.

Explore →

Research, Standards & Evaluation

Practical methods for assessing generative and agentic AI systems.

OWASP GenAI AI Red Teaming Manual

Core contributor to methodologies for evaluating safety, robustness, trustworthiness, and resilience.

Cloud Security Alliance

Contributor to agentic AI and red teaming initiatives.

OWASP Agentic Top 10

Contributor to emerging guidance on autonomous and agentic AI risks.

The Architecture of Intelligence

Author of a framework for understanding AI agent systems and governance implications.

Speaking & Teaching

Training across professional communities, conferences, universities, and AI ecosystem programs.

600+ professionals trained in AI Security, Safety, Governance, and Agentic AI.

Workshops and talks across ISACA, UCSC Silicon Valley Extension, ISC2, NJSECON, Society for Digital Mental Health, BioNJ, NJ AI Hub, Plug and Play, universities, podcasts, and industry events.