About

The Researcher

Ali Akarma
Madinah, Saudi Arabia
Islamic University of Madinah
12 Publications

Research Areas

Agentic AI & Autonomous Systems in High-Stakes Environments
AI Safety & Alignment of Large Language Models
Prompt Injection, Jailbreaks & Adversarial Misuse
Governance, Oversight & Constitutional AI
Secure AI for Cybersecurity & Critical Infrastructure
Digital Twins & Smart City AI

Biography

AI researcher focused on trustworthy and safety-aligned agentic systems, with peer-reviewed publications spanning AI architectures, governance, and real-world deployment contexts. Experienced in designing constrained agent pipelines, analyzing failure modes, and studying robustness in high-stakes environments. Preparing for graduate research in reliable machine learning systems.

Currently pursuing a B.S. in Information Technology on a merit-based fully funded scholarship, Ali has authored or co-authored papers across IEEE conferences, Springer, PLOS ONE, and MDPI — spanning agentic architectures, digital twins, AI governance, and safety-critical deployment. His research is motivated by a conviction that powerful AI systems must be understandable, governable, and safe by design.

Education

B.S. Information Technology

Islamic University of Madinah

Saudi Arabia

2023 – 2027Current
  • Merit-based fully funded scholarship
  • Undergraduate researcher in agentic AI systems and AI governance

Research Experience

Undergraduate AI Researcher

Islamic University of Madinah

Jan 2025 – Present
  • Designed agentic AI systems with explicit safety constraints, restricted action spaces, and policy-aligned decision rules.
  • Conducted failure-mode and misuse analyses for autonomous systems, focusing on adversarial inputs, unintended behaviors, and oversight gaps.
  • Explored LLM-based components within agentic pipelines: prompt design, orchestration, and robustness considerations.
  • Investigated governance mechanisms for preventing unsafe behaviors triggered by untrusted or adversarial data sources.
  • Analyzed system trustworthiness in scenarios analogous to log analysis, automated triage, and decision support pipelines.

Awards & Certifications

Merit-Based Fully Funded Bachelor's Scholarship

Islamic University of Madinah

2023

Certificate of Appreciation

ICBDT 2025

2025

Mindware: Critical Thinking

University of Michigan

2024

Computational Thinking

University of Michigan

2024