name: Eyal
education:
university: Ben-Gurion University of the Negev
degree: B.Sc. Data Science (3rd Year)
achievements: Faculty's Honours List x2 (2024,2025)
research:
paper: "GAVEL: Towards rule-based safety through activation monitoring"
venue: ICLR 2026 (co-author)
topic: LLM safety via activation-based monitoring
currently: Research Assistant @ Offensive AI Lab
fun_fact: This GitHub is way cooler than my Google DriveGAVEL: Towards rule-based safety through activation monitoring
Co-author — A framework for compositional, rule-based activation safety in LLMs. Improves precision, supports domain customization, and enables auditable AI governance. Accepted at ICLR 2026.
📂 InformationRetrievalProj2025-26
├── 🔎 BM25 + Cosine Similarity ranking
├── 📊 PageRank & PageView integration
├── ☁️ Deployed on Google Colab & GCP
├── ⚙️ Grid search hyperparameter tuning
└── 📚 Full English Wikipedia corpus (up to Aug 2021)
📂 VisualizationProject
├── 📈 Interactive Streamlit dashboard
├── 🗺️ GeoJSON district map visualization
├── 📉 Plotly & Matplotlib charts
├── 🔄 Multi-page data exploration
└── 🌐 Live app: streamlit.app
📂 SafariSynchronisation
├── 🧵 Thread handling & concurrency in C#
├── 🔒 Async tasks, locks, mutexes, semaphores
├── 🛑 Cancellation Token patterns
└── 📝 Developed Backend synchronisation methods and authored project report



