I am a Software Engineering undergraduate majoring in Cybersecurity at Daffodil International University, Bangladesh.
My primary interests include Penetration Testing, Bug Hunting, Vulnerability Assessment, Web Application Security, Network Security, and AI Security Research.
I am passionate about identifying security weaknesses, analyzing attack surfaces, understanding adversarial techniques, and developing practical security solutions.
I am also exploring the intersection of Artificial Intelligence and Cybersecurity, with a particular research focus on Prompt Injection Detection, Defense, and Deep Security Analysis in Large Language Models (LLMs).
AI-Powered Prompt Injection Detection, Defense, and Deep Analysis in Large Language Models
My research interests include:
- Direct and Indirect Prompt Injection
- Jailbreak and Adversarial Prompt Analysis
- Prompt Injection Detection Techniques
- Attack Classification and Risk Assessment
- LLM Security Evaluation and Benchmarking
- Defensive Mechanisms and Security Architecture
- AI-driven Security Analysis
PromptGuard AI
An independent research and development project focused on building an AI-powered system to detect, classify, assess, and defend against prompt injection attacks targeting LLM-based applications.
- Penetration Testing
- Bug Hunting
- Web Application Security
- Vulnerability Assessment
- Network Reconnaissance
- Security Testing and Automation
- Degree: B.Sc. in Software Engineering
- Major: Cybersecurity
- University: Daffodil International University
- Expected Graduation: April 2027
Cybersecurity Security Analysis Vulnerability Assessment |
Offensive Security Penetration Testing Bug Hunting |
AI Security LLM Security Prompt Injection Defense |
| Research Domain | Areas of Interest |
|---|---|
| AI & LLM Security | Prompt Injection, Jailbreak Detection, LLM Defense |
| Web Security | OWASP Top 10, Web Vulnerability Assessment |
| Penetration Testing | Reconnaissance, Enumeration, Security Testing |
| Network Security | Network Scanning, Service Enumeration |
| Security Automation | Python-based Security Tools |
| Digital Forensics | Security Investigation and Evidence Analysis |
Additional technologies planned for PromptGuard AI
These are part of the project's intended technology stack and should not be interpreted as completed implementation experience.
- FastAPI
- React
- PostgreSQL
- Hugging Face Transformers
- Machine Learning and Deep Learning
Project Overview
PromptGuard AI is an independent research and development project focused on detecting, classifying, assessing, and defending against prompt injection attacks targeting Large Language Models and LLM-powered applications.
Core Research & Development Modules
| Module | Objective |
|---|---|
| Detection | Identify malicious prompt injection attempts |
| Classification | Categorize different attack types |
| Risk Assessment | Evaluate potential security risks |
| Defense | Explore defensive prompt security mechanisms |
| Logging | Record and analyze security events |
| REST API | Provide programmatic security functionality |
| Dashboard | Visualize security findings |
| Evaluation | Benchmark detection and defense performance |
Planned Technology Stack
Python Machine Learning Hugging Face Transformers FastAPI React PostgreSQL
Repository and public implementation will be linked after the project is ready for release.
- Singleton Design Pattern — Java-based software design pattern implementation.
- Digital Supply Tracker — Digital supply chain tracking and UI/UX design project.
| Program / Platform | Focus |
|---|---|
| Ostad | Cyber Security & Ethical Hacking Career Track |
| Hack The Box | Hands-on Security Learning |
| CodeSignal | Programming Practice |
Open to cybersecurity discussions, research collaboration, and professional opportunities.
