top of page
Featured in the Journal of Computer Information Systems

Executive Summary: This research introduces the LLM Scalability Risk Index (LSRI), a parametric framework to stress-test Agentic-AI in security-critical environments. It further proposes a model-supply-chain framework to establish a verifiable root of trust throughout the model lifecycle.

https://doi.org/10.1080/08874417.2026.2624670

  • Keywords: AI Governance, Agentic-AI, LSRI, Model Supply Chain Security, Cybersecurity, Dual-use AI.

After months of rigorous peer-review, our latest paper that s the result f our collaboration with Google, Mountain View, CA, got published yesterday in Journal of Computer Information Systems by Taylor & Francis with  

  • Impact Factor (2024): 4.2

  • 5-Year Impact Factor: 3.9

  • SCImago Journal Rank (SJR): 0.882 (2024)

  • Scopus CiteScore: 9.3 (2024)

  • Publisher: Taylor & Francis 

Link to the page for more details: Click Here

Download full paper: LLM Scalability Risk for Agentic-AI and Model Supply Chain Security (PDF)

https://www.tandfonline.com/doi/citedby/10.1080/08874417.2026.2624670?scroll=top&needAccess=true

Ahi, K., Agrawal, V., & Valizadeh, S. (2026). LLM Scalability Risk for Agentic-AI and Model Supply Chain Security. Journal of Computer Information Systems, 1–17. https://doi.org/10.1080/08874417.2026.2624670

SPIE.jpg

LLMs and LVMs for agentic AI: a GPU-accelerated multimodal system architecture for RAG-grounded, explainable, and adaptive intelligence

Proceedings Volume 13687, Photomask Technology 2025; 136871R (2025) https://doi.org/10.1117/12.3078485
Event: SPIE Photomask Technology + EUV Lithography, 2025, Monterey, California, United States​

Citation Download Citation

Kiarash Ahi, Chih-Hung Hsieh, and Germain Fenger "LLMs and LVMs for agentic AI: a GPU-accelerated multimodal system architecture for RAG-grounded, explainable, and adaptive intelligence", Proc. SPIE 13687, Photomask Technology 2025, 136871R (6 November 2025); https://doi.org/10.1117/12.3078485

TY  - CONF
TI  - LLMs and LVMs for agentic AI: a GPU-accelerated multimodal system architecture for RAG-grounded, explainable, and adaptive intelligence
AU  - Kiarash Ahi
AU  - Chih-Hung Hsieh
AU  - Germain Fenger
T2  - Proc.SPIE
VL  - 13687
SP  - 136871R

UR  - https://doi.org/10.1117/12.3078485
PY  - 2025/11/6
DO  - 10.1117/12.3078485
ER  - 
 

In Collaboration with Siemens

Rigorously Peer Reviewed

Recommended by Larry Melvin of Synopsys Inc

Download full paper: LLMs and LVMs for agentic AI: a GPU-accelerated multimodal system architecture for RAG-grounded, explainable, and adaptive intelligence (PDF)

bottom of page