Welcome to Moosec!
Our research group is located at the CISPA Helmholtz Center for Information Security in Saarbrücken, Germany, where we work at the intersection of machine learning 
security.
We attack machine learning systems, defend them, and ask the questions about what makes them fail. This often means going beyond the model — into pre-processing, post-processing, and the hardware and software stack beneath. We also build on modern LLMs and agent systems to push forward on practical security problems: vulnerability analysis, fuzzing, malware classification.
If you are interested in our research, take a look at our recent publications or reach out via mail.
News
Huzaifa puts down roots and starts his PhD 🌱, and a new moose joins the herd — welcome our new Research Assistant Rishabh! 🦌
PINE 2026 🌲 kicks off at Saarland University. A fresh cohort taking on ML security challenges in our hands-on seminar.
LLM-based Vulnerability Discovery at ICSE 2026. Surprisingly, a simple code metrics classifier can match the performance of state-of-the-art LLMs on classification of vulnerable code functions.
SaTML in Munich: great talks, nice conversations, and 🥨 way too many pretzels.
Guest lecture at Reykjavik University 🌌. Joined Giovanni Apruzzese's group remotely to talk about ML security.
more
We presented Chasing Shadows at NDSS 2026, looking at why so much LLM security research keeps going sideways. Spoiler: it's complicated.
Happy to contribute a challenge and attend the CISPA European Hackathon Championship in Stockholm 🇸🇪. Great energy, terrible sleep schedules.
Guest lecture at TU Vienna ☕. Great to visit Daniel Arp and talk about ML security research.
A first moose finds the forest 🌲. Huzaifa joins as a research intern. Welcome!
moosec is officially live at CISPA 🚀 Excited to get started!
Team
Publications
2026
Network and Distributed System Security Symposium (NDSS)
IEEE/ACM International Conference on Software Engineering (ICSE)
Detection of Intrusions and Malware & Vulnerability Assessment (DIMVA)
ACM Asia Conference on Computer and Communications Security (ASIACCS)
2025
ACM Conference on Computer and Communications Security (CCS) ★ Distinguished Paper Award
International Conference on Machine Learning (ICML)
USENIX Security Symposium
IEEE Conference on Secure and Trustworthy Machine Learning (SaTML)
Detection of Intrusions and Malware & Vulnerability Assessment (DIMVA)
2024
IEEE Symposium on Security and Privacy (S&P)
ACM Asia Conference on Computer and Communications Security (ASIACCS)
ACM Asia Conference on Computer and Communications Security (ASIACCS)
2023
Dissertation ★ Faculty Award for Outstanding Achievement
USENIX Security Symposium
IEEE Conference on Secure and Trustworthy Machine Learning (SaTML)
Network and Distributed System Security Symposium (NDSS)
European Symposium on Artificial Neural Networks (ESANN)
2022
Annual International Cryptology Conference (CRYPTO)
2021
2020
International Conference on Machine Learning (ICML)
Annual Computer Security Applications Conference (ACSAC)
Teaching
Research Problems in Machine Learning and Security ● SS 2026
Students work in small teams on hands-on challenges spanning both attacks against ML systems and security applications of machine learning. Challenges run as Kaggle-style competitions with a shared scoreboard, mixing independent research with collaborative exploration.
Past Teaching
Full teaching history at eisenhofer.me.
Open Positions
We currently have no open positions. That said, we are always open to hearing from motivated people — if our research resonates with you, feel free to send an initiative application.
Send your CV, transcripts, and a short note to eisenhofer@cispa.de.