Cybersecurity Specialist
Leading cybersecurity initiatives focusing on AI-powered threat analysis, privacy assessments, and risk management strategies for critical systems including automotive cybersecurity and digital identity solutions.
Majid Mollaeefar
Cybersecurity Researcher | AI & Privacy
About Me
Professional Summary
Cybersecurity researcher with a Ph.D. in Computer Science and Systems Engineering from the University of Genova, Italy. With 6+ years of experience in cybersecurity risk frameworks, threat modeling, and privacy engineering, I have more recently expanded my research into AI-powered security by designing LLM-based tools for threat modeling, privacy risk analysis, and Cyber Threat Intelligence as a postdoctoral researcher at Fondazione Bruno Kessler (FBK), Trento. Current research spans multi-agent AI systems, LLM benchmarking for security tasks, RAG-based approaches, and risk frameworks for trustworthy AI, with results published in high-quality venues including PETS, JISA, and FGCS.
0
Publications
6+
Years Experience
0
Tools
0
Students Supervised
Core Expertise
Risk Assessment & Management
Privacy Protection & GDPR
Threat Modeling
AI-Powered Security
DPIA & Compliance
Python, Java, JavaScript
Large Language Models
Multi-Agent AI Systems
LLM Benchmarking & Evaluation
Prompt Engineering
LINDDUN Privacy Framework
Professional Experience
Cybersecurity Researcher (Ph.D. Career)
Developed innovative methodologies for multi-stakeholder risk assessment, conducted Data Protection Impact Assessments (DPIA), and created AI-powered tools for automated threat analysis and privacy risk modeling.
Junior Information Security Specialist
Conducted comprehensive gap analyses to achieve ISO 27001 compliance, drafted and implemented ISMS documentation, and delivered training programs to enhance organizational security awareness.
Education
Ph.D. in Cybersecurity
Dissertation: "Automating the Quantification and Mitigation of Risks for Multiple Stakeholders"
View ThesisPublications
2026
Benchmarking the Effectiveness of Multi-agent LLMs in Collaborative Privacy Threat Modeling with LINDDUN GO
Journal of Information Security and Applications (JISA) Journal
2026
Chatbot Confessions: Large-Scale Analysis of Private Data Disclosure in Shared AI Chatbot Conversations
Privacy Enhancing Technologies Symposium (PETS) Conference
2025
A comparative benchmark study of LLM-based threat elicitation tools
Journal of Future Generation Computer Systems (FGCS) Journal
2025
Multi-entity Control-based Risk Assessment: A European Digital Identity Wallet Use Case
The 30th Nordic Conference on Secure IT Systems (Nordsec25) Conference
2025
PILLAR: LINDDUN Privacy Threat Modeling using LLMs
International Workshop on Privacy Engineering (IWPE'25) Workshop
2024
Modeling and Assessing Coercion Threats in Electronic Voting
19th International Conference on Risks and Security of Internet and Systems (CRiSIS) ConferenceBest Paper 🏆
2024
Protecting Digital Identity Wallet: A Threat Model in the Age of eIDAS 2.0
19th International Conference on Risks and Security of Internet and Systems (CRiSIS) Conference
2024
A Risk-based Approach to Trustworthy AI Systems for Judicial Procedures
4th Ital-AI Workshop AI Responsabile e Affidabile Workshop
2023
Identifying and Quantifying Trade-offs in Multi-Stakeholder Risk Evaluation with Applications to the Data Protection Impact Assessment of the GDPR
Journal of Computers & Security Journal
2020
Multi-Stakeholder Cybersecurity Risk Assessment for Data Protection
17th International Conference on Security and Cryptography (SECRYPT 2020) Conference
2017
A novel encryption scheme for colored image based on high level chaotic maps
Journal of Multimedia Tools and Applications Journal
2017
A novel method for digital image steganography based on a new three-dimensional chaotic map
Journal of Multimedia Tools and Applications Journal
2017
An improved method for digital image fragile watermarking based on chaotic maps
Journal of Multimedia Tools and Applications Journal
2016
A novel method for image encryption using chaotic maps
3rd International Conference on Applied Research in Computer and Information Technology Conference
2015
An improved method for image encryption based on high level chaotic maps and improved gravity model
International Congress on Technology, Communication and Knowledge (ICTCK) Conference
Developed Tools
AutoSecGPT
Description: An AI-powered tool designed to automate threat modeling for automotive applications, aligned with ISO/SAE 21434 standard. AutoSecGPT supports security teams by facilitating the entire cybersecurity engineering process—from threat identification to risk assessment.
Key Features:
- Works with different LLMs such as OpenAI to analyze threat scenarios
- Identifies security threats specific to automotive systems
- Generates detailed threat scenarios and descriptions to facilitate risk understanding
- Visualizes threat scenarios through attack graphs
- Conducts likelihood and impact assessments to prioritize threats
- Supports Threat Analysis and Risk Assessment (TARA) as per ISO/SAE 21434
PILLAR
Description: Privacy risk Identification with LINDDUN and LLM Analysis Report (PILLAR) is an AI-powered tool developed to automate and enhance privacy threat modeling. Built on the LINDDUN framework, PILLAR leverages Large Language Models (LLMs) to automatically generate Data Flow Diagrams (DFDs), identify privacy threats, and prioritize risks.
Key Features:
- Automates privacy threat modeling using LINDDUN and LLMs
- Generates DFDs from natural language descriptions
- Simulates multi-agent collaboration for comprehensive threat analysis
- Assesses the impact of identified privacy threats
- Suggests control measures based on privacy patterns
- Generates comprehensive reports on privacy threat modeling
Multi-Stakeholder Risk Assessment Tool
Description: A comprehensive risk assessment tool that enables risk analysts to perform risk evaluation in a multi-stakeholder manner for a given system. The tool addresses the complexity of balancing different stakeholders' interests in risk management decisions.
Purpose:
- Evaluate and quantify risk levels for all involved stakeholders
- Solve the risk minimization problem as a multi-objective optimization challenge
- Identify trade-offs between conflicting stakeholder interests
- Support decision-making with quantitative risk data
DIWAR
Description: Digital Identity Wallet Analysis and Risk assessment (DIWAR) is a control-based risk assessment tool specifically designed for Digital Identity Wallet ecosystems. It integrates DREAD factors with entity-specific control attribution to quantitatively evaluate threats and prescribe targeted mitigations aligned with each entity's responsibilities.
Key Features:
- Entity-specific risk assessment for Issuers, Verifiers, and Wallet Providers
- Control-based evaluation using the (modified) DREAD model (Damage, Reproducibility, Exploitability, Affected users, Detectability)
- Multi-level control implementation (Basic, Intermediate, Advanced)
- Qualitative risk matrix with color-coded severity levels
- Interactive threat and control visualization
- Role-specific security control recommendations