Daniel Alfasi
I'm an AI researcher, motivated by solving real-life cybersecurity problems. 🤖 🔒 💡
TBD • 2025 - Present
Starting a new adventure 🚀
CyberArk • 2023 - 2025
Worked on various cybersecurity projects focused on attack detection. Developed anomaly detection models to detect both low- and high-rate attacks. Researched solutions for modeling user behavior and user-asset interactions using Graph Neural Networks (GNNs) and Natural Language Processing (NLP).
CyberArk • 2020 - 2023
Worked on a large-scale detection system for privileged users. Designed and developed a data-driven architecture using AWS Lambdas.
Surecomp Innovation Lab • 2019 - 2020
Conducted multidisciplinary research on emerging development technologies. Focused on securing client transactions, mutual authentication between endpoints, and developed POCs for fraud detection, KYC, and transaction security.
Illusive Networks (acquired by Proofpoint) • 2018 - 2019
First cybersecurity role in the industry. Worked on the discovery of organizational assets (crown jewels), shadow admins, and securing the organizational attack surface.
Perfecto (acquired by Perforce) • 2016 - 2018
Worked on SaaS platform over AWS.
Reichman University • 2020 - 2023
Magna Cum Laude
Thesis: Detecting links between security entities (CVEs, CPEs, and CWEs) using Graph Neural Networks (GNNs) and Natural Language Processing (NLP).
Supervisors: P.hD. Tal Shapira and Prof. Anat Bremler-Barr
Research conducted at the DEEPNESS lab of Tel-Aviv University.
My research was supported by Red Hat and Google Cloud Research.
Ariel University • 2015 - 2017
Specialized coursework in cryptography, probability theory, and cybersecurity labs.
Reichman University • 2020 - Present
Active member of Momentum, Reichman University's startup accelerator, which is part of Israel's leading university for entrepreneurship.
November 14-17, 2024
Over 500 participants from 48 countries, and a prize pool of 26k USD.
Mentored at AGENT CRAFT HACKATHON - sponsored by LangChain, advising on agent architectures and evaluation methods.
Authors: Daniel Alfasi, Tal Shapira, and Anat Bremler-Barr
GNNet Workshop @ CoNEXT 2024 (University of California, Los Angeles).Inductive Link Prediction in cybersecurity Knowledge Graph using ULTRA (Galkin et al.) and Large Language Models (LLMs).
Authors: Daniel Alfasi, Tal Shapira, and Anat Bremler-Barr
SYSTOR '23 (Technion Institute of Technology, Haifa, Israel)Transductive Link Prediction in cybersecurity Knowledge Graph using Graph Neural Networks (RotatE) and Large Language Models (LLMs).
Medium • 2024
Technical details on how we integrated GNN with LLM to enrich entity-level features for a link prediction task.
Medium • 2023
Exploring new ways to understand the scope of vulnerabilities using LLMs and Knowledge Graphs. Specifically, discovering new vulnerable products that have not been reported yet.
This project is an educational project for implementing computational graphs in Golang, inspired by the nn-zero-to-hero by Andrej Karpathy.