Kshitiz Aryal karyal@uno:~/research-portfolio
whoami faculty_profile.txt

Assistant Professor | University of Nebraska Omaha

Kshitiz Aryal

cat profile.md

I am an Assistant Professor in the School of Interdisciplinary Informatics at the University of Nebraska Omaha. My research focuses on securing AI systems that act, remember, communicate, and operate in cyber environments. I work at the intersection of agentic AI security, adversarial machine learning, malware analysis, binary security, and autonomous red teaming.

Our group studies how modern AI systems fail under adversarial pressure, how those failures can be measured through reproducible benchmarks, and how robust defenses and autonomous security tools can be built for real-world cyber settings.

tail -f news.log

News

  • Mar 2026Paper accepted at The 2nd International Conference on Artificial Intelligence Systems (AIS 2026).
  • Dec 2025Paper accepted at IEEE 6th IEEE International Conference on AI in Cybersecurity (ICAIC) 2026.
  • Nov 2025Paper accepted at IEEE BigData 2025.
  • Sep 2025Paper accepted at Computers and Security.
  • Aug 2025Joined UNO as Assistant Professor.
  • May 2025Defended Ph.D. at Tennessee Tech.
  • May 2024Poster at RSA Conference 2024.
tree research/

Research portfolio

A unified research effort in agentic security and adversarial machine learning: discover attack surfaces, formalize threat models, build benchmarks, generate attacks, evaluate defenses, and automate red teaming.

01. Agentic AI attack-surface discovery

Security analysis of agentic AI systems.

  • Orchestration-layer security: attack surface by frameworks.
  • Memory, RAG, and LTMS security: compromise in agent memory.
  • MCP security: risks in agent-tool communication.
  • A2A security: vulnerabilities in inter-agent discovery.
  • LLM agents
  • Agent Tools
  • Agent Memory
  • Agent Protocols

02. Autonomous cybersecurity agents

Building agentic systems that can support the vulnerability lifecycle.

  • Automated vulnerability scanning: agents for discovery and triage
  • Patch generation and remediation: agent-assisted patching, regression checks, and reporting.
  • Red Teaming Agents
  • patching
  • agentic security

03. Adversarial ML for malware and binary security

Attacking, evaluating, and defending ML-based malware detectors and binary-analysis systems under realistic adversarial constraints.

  • Malware detection: detection, adversarial attacks, and defense
  • Binary analysis: reverse engineering, and XAI for security.
  • malware ML
  • binary rewriting
  • reverse engineering
  • XAI

04. Autonomous ML and agentic red-teaming platform

  • Agentic framework: LLM-powered red-team agents for attack planning, benchmark execution, and evidence collection.
  • Unified reporting: reproducible reports for agentic systems, malware ML, binary attacks, and cyber-agent evaluations.
  • red teaming
  • evaluation
  • defenses
cat research_foundations.md

Research foundations and continuing lines

Explainability-guided adversarial evasion

Uses SHAP values to attribute CNN malware detections to PE regions, enabling fine-grained perturbation strategies and improved evasion effectiveness.

  • SHAP
  • Windows PE
  • Adversarial evasion
  • XAI

Intra-section code cave injection

A code-cave and loader approach to hide adversarial perturbations across PE files while preserving malware functionality and improving evasion.

  • Reverse engineering
  • Code caves
  • FGSM

GenAI in cybersecurity and privacy

Evaluates misuse and defensive potential of generative AI tools, including security feature comparisons, privacy implications, and ethical risks.

  • LLMs
  • Privacy
  • Security evaluation

Poisoning and model hardening

Studies label-flip poisoning on ML-based malware detectors and develops practical hardening strategies for robust security models.

  • Data poisoning
  • Label flipping
grep -i publications.json

Publications

cat teaching.log

Teaching

  • Instructor: Malware Analysis (CYBR445/CYBR8456), Spring 2026
  • Instructor: Cyber Investigation (CYBR8490), Fall 2025
  • Instructor: Introduction to Problem Solving & Programming (CSC1300), Spring 2025
  • TA: AI-Assisted Malware Analysis (CSC7570), Fall 2022/2023/2024
  • TA: Cloud Security (CSC6570), Spring 2023/2024
  • TA: NSA Cloud Security Workshop & Faculty Development Workshop, Summer 2023
cat mentoring.md

Mentoring

  • Advising and co-advising graduate and undergraduate students in AI security research.
  • Co-mentored two M.S. students to complete thesis/projects; supporting additional students in related areas.
  • Workshop mentor: "Introducing ThreatGPT: The Malicious Sibling of ChatGPT" at WiCyS 2024.
cat service.txt

Service

  • Program Committee: 31st European Symposium on Research in Computer Security (ESORICS) 2026
  • Proceedings and Web Chair: ACM SaT-CPS 2026
  • Program Committee: ACM SaT-CPS 2025
  • Reviewer: IEEE TIFS, IEEE TDSC, IEEE TAI, IEEE TR, IEEE Access, ACM TOPS, ACM TSE
  • Reviewer/Subreviewer: CODASPY 2022-2024, DBSec 2022, DFRWS EU 2024
  • Volunteer: WiCyS 2023, WiCyS 2024
ssh contact

Contact

The fastest way to reach me is by email. I am happy to discuss research collaborations, student opportunities, and speaking engagements.

mail message.txt

Send a message