CV
الْحَمْدُ لِلَّهِ الَّذِي رَزَقَنِي بِغَيْرِ حَوْلٍ مِنِّي وَلَا قُوَّةٍ وَلَطَفَ بِي فِي قَضَائِهِ وَقَدَرِهِ
General Information
| Full Name | Adel Samir ElZemity |
| ae455@kent.ac.uk | |
| Address | Canterbury, CT1 1DS, United Kingdom |
| Languages | Arabic (Native), English (Fluent), German (Intermediate) |
| Summary | A Doctoral Researcher in Computer Science at the University of Kent, UK, with a background as a software engineer. My research focuses on the safety and security of Large Language Models (LLMs) and their applications in cyber security. I am passionate about AI and quickly adapt to new frameworks and technologies. I am particularly excited about advancing AI safety and security. |
| Links |
Education
-
2023 - now United Kingdom
PhD in Computer Science
University of Kent (UKC) - Focus on AI safety and cyber security
- Full-Merit scholarship
-
2021 USA
Exchange Semester (Global UGRAD) in Computer Engineering
Fayetteville State University (UNCFSU) - GPA (4.00/4.00) with Honors (President's List)
- Full-Merit scholarship
-
2021 Latvia
Exchange Semester (Erasmus+) in Computer Engineering
Riga Technical University (RTU) - GPA (4.00/4.00) with Honors (President's List)
- Full-Merit scholarship
-
2018 - 2023 Egypt
Bachelor's of Science in Computer Engineering
Nile University (NU) - GPA (3.9/4.0) with Honors (President's List, Dean's List)
- Full Merit Scholarship
Work Experience
-
2023 - now United Kingdom
Doctoral Researcher
University of Kent - Leading research on LLM safety and security, developing evaluation frameworks to assess risks in fine-tuned models
- Building machine learning pipelines to train and evaluate LLMs on cyber security tasks, with focus on safety alignment
- Part of the "Countering HArms caused by Ransomware on the Internet of Things (CHARIOT)" project funded by EPSRC in the UK
-
2023 - now United Kingdom
Teaching Assistant
University of Kent - Teaching assistant for undergraduate and postgraduate courses in Computer Science
- Supporting students in programming, algorithms, and software development
- Conducting lab sessions and providing academic guidance
-
2022 - 2023 Spain
Machine Learning Engineer
National Cancer Research Center - Tested and optimised supervised ML models to detect metal binding sites in proteomes
- Participated in the UniProt Machine Learning challenge 2022
- Created python pipeline for exploring, cleaning, and filtering datasets to improve accuracy and efficiency
-
2021 - 2022 USA
Software Engineer
Intelligent Systems Lab (ISL) - Applied expertise in deep learning to design and implement an architecture for robot's ZED Camera
- Improved efficiency of object detection on the moon by 2%
- Configured 24 Linux-based robots using RaspberryPi and Jetson Nano using ROS and Python
Publications
-
2025 Analysing Safety Risks in LLMs Fine-Tuned with Pseudo-Malicious Cyber Security Data
- Published in ArXiv
- Adel ElZemity, Budi Arief, Shujun Li
-
2025 CyberLLMInstruct A New Dataset for Analysing Safety of Fine-Tuned LLMs Using Cyber Security Data
- Published as Dataset Paper in ArXiv
- Adel ElZemity, Budi Arief, Shujun Li
-
2024 Privacy Threats and Countermeasures in Federated Learning for Internet of Things
- Published in IEEE International Conference on Internet of Things (iThings)
- Adel ElZemity, Budi Arief
-
2023 A Comparative Analysis of Time Series Transformers and Alternative Deep Learning Models for SSVEP Classification
- Published in International Conference on Model and Data Engineering
- Heba Ali, Adel ElZemity, Amir E Oghostinos, Sahar Selim
-
2023 A Transformer-Based Deep Learning Architecture for Accurate Intracranial Hemorrhage Detection and Classification
- Published in IEEE International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT)
- ElZemity et al.
-
2020 Wastewater Treatment Model with Smart Irrigation Utilizing PID Control
- Published in IEEE Novel Intelligent and Leading Emerging Sciences Conference (NILES)
- ElZemity et al.
-
2019 Interfacial Modification of Perovskite Solar Cell Using ZnO Electron Injection Layer with PDMS as Antireflective Coating
- Published in IEEE Novel Intelligent and Leading Emerging Sciences Conference (NILES)
- Mohamed K. Othman, Adel ElZemity, Mohamed K. Rawash, Hazem A. Taha, Shorouk Alalem, Maryam El-Fdaly, Yasser M. El-Batawy
Technical Skills
-
Development & Tools
- Python, Shell, HTML, CSS, JavaScript, SQL, TypeScript, Swift
- Flutter, React, Next.js, Wordpress
- VSCode, Cursor, Terminal, XCode, PyCharm
- Google Colab, Kaggle, Hugging Face
- LaTex, Markdown, Word, Excel, PowerPoint, PDF, Overleaf
- AWS, Azure, GCP, Vercel, GitHub Pages, Cloudflare
- Git, GitHub, GitLab, Jupyter Notebook
- Docker, Kubernetes, Virtual Machines, Virtualization
- Linux, Windows, MacOS
- API, REST, Webhooks, WebSockets
- CI/CD, DevOps, Cloud Computing, Cloud Infrastructure
- Data Science, Data Engineering, Data Analysis, Data Visualization
-
Artificial Intelligence
- Machine Learning, Deep Learning, Computer Vision, Natural Language Processing
- Time Series Analysis, Fine-tuning LLMs, Retrieval-Augmented Generation (RAG)
- Federated Learning, Federated Learning for IoT, Flower framework
- TensorFlow, PyTorch, Keras, Scikit-learn, Pandas, NumPy, Matplotlib, Seaborn
- ChatGPT, Claude, Gemini, Llama, Qwen, Mistral, Gemma, DeepSeek
-
Cyber Security
- Network architecture & security (TCP/IP, VLANs, routing, switching, VPNs)
- IoT Security (LoRaWAN, Thread, Zigbee, Bluetooth, Wi-Fi)
- Access control systems (RFID, NFC, Biometrics)
- Capture the Flag (CTF)
- Malware analysis (Malware, Ransomware, Botnet, etc.)
- AI Security, attacks, and defenses (LLM, RAG, Federated Learning, etc.)
- Privacy and Data Protection (GDPR, CCPA, HIPAA, etc.)
- Encryption and Decryption (AES, RSA, SHA, etc.)
Open Source Projects
-
2025 CyberLLMInstruct
- A comprehensive dataset and framework for analysing safety of fine-tuned LLMs using cyber security data. Features include dataset creation pipeline, fine-tuning support for multiple LLMs (Phi-3, Mistral, Qwen, Llama, Gemma), and evaluation tools.
-
2025 Canterbury Mosque Prayer Times Extension
- A Chrome extension that displays daily prayer times from Canterbury Mosque with notifications and adhan audio. Features include automatic updates, customisable notifications, and a user-friendly interface.
-
2024 RPL Attack Implementations for Contiki-NG and COOJA
- A comprehensive implementation of RPL (Routing Protocol for Low-power and Lossy Networks) attacks using Contiki-NG and COOJA. Features include implementation of various attacks (Selective Forwarding, Sinkhole, Version Number, DIS Flooding, and Sybil attacks), simulation scripts, and analysis tools.