cv
Ali El Husseini's CV.
Contact Information
| Name | Ali El Husseini |
| Professional Title | PhD Candidate, National University of Singapore |
| elhusseiniali@u.nus.edu |
Professional Summary
PhD candidate in the Trustworthy and Secure Software group at the National University of Singapore, advised by Prof. Abhik Roychoudhury. My research focuses on improving the security of large language models (and the agentic systems built on top of them) by understanding and analyzing how these models operate.
Experience
-
2022 - 2022 Paris, France
Research Intern
INRIA Paris
Worked on an abstract domain to represent data transformations inside Jupyter notebooks, implemented within LiSA (https://github.com/UniVE-SSV/pylisa/), to study the safety and sensitivity of notebooks to changes in data shapes and the soundness of results. Supervised by Dr. Caterina Urban.
-
2021 - 2021 France
Research Intern
Commissariat à l'Énergie Atomique (CEA)
Used abstract interpretation to verify formal properties (e.g., robustness) of neural networks, focusing on the soundness of the numerical computations required by the analysis on GPUs. Supervised by Dr. Zakaria Chihani and Augustin Lemesle.
-
2020 - 2020 France
Research Intern
Centre National de la Recherche Scientifique (CNRS)
Developed a modern Python library to handle Simple Regular Expressions (see https://doi.org/10.1023/B:FORM.0000033962.51898.1a), to facilitate the analysis of lossy channel systems and verification of formal properties. Supervised by Dr. Philippe Schnoebelen. Code: https://github.com/elhusseiniali/sre.
Education
-
2023 - present Singapore
PhD Candidate, Trustworthy and Secure Software group
National University of Singapore
Computer Science
- Advised by Prof. Abhik Roychoudhury.
- Research focus: improving the security of large language models and LLM-based agent systems.
-
2019 - 2022 Paris, France
M1 and M2, Parisian Masters of Research in Computer Science (MPRI)
École Normale Supérieure Paris-Saclay
Computer Science
-
2014 - 2019 Beirut, Lebanon
BS in Computer Science, BA in Philosophy
American University of Beirut
Computer Science and Philosophy
Publications
-
2026 Repairing LLM Executions for Secure Automatic Programming
International Conference on Software Engineering (ICSE)
Proposed a novel approach to identify and repair internal representations that cause a large language model to produce vulnerable code, via a tool that monitors the LLM at run-time and intervenes to improve the security of the generated code.
-
2025 Detecting Python Malware in the Software Supply Chain with Program Analysis
International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP)
Proposed a program-analysis-based approach to detect supply chain attacks in the Python ecosystem, combining AST-level differential analysis, CodeQL-based behavioral detection, and transitive dependency tracking to identify malicious Python packages.