Programming Languages and Verification
Computer Science
1 Career
- Since 2025 Assistant Professor, Institute of Science and Technology Austria (ISTA)
- 2024 Postdoctoral researcher, ETH Zürich, Switzerland
- 2019 – 2023 PhD, Max Planck Institute for Software Systems, Germany
1.1 Selected Distinctions
- 2026 Organizer of the Iris Workshop 2026, ISTA / June 8–12, 2026
- Otto-Hahn-Medal, Runner-Up for Informatics Europe 2024 Best Dissertation Award, and Dr. Eduard Martin Prize for PhD thesis on Automated and Foundational Verification of Low-Level Programs
- Distinguished Paper Award at PLDI’25, POPL’23, POPL’22, PLDI’21, and USENIX Security’19
- 2020 – 2023 Google PhD Fellowship
1.2 Selected Grants
- N/A
1.3 Tech Transfer (Patents, Spin-Offs)
N/A
1.4 Selected Publications
- N. Mück, A. L. Georges, D. Dreyer, D. Garg, and M. Sammler, “Endangered by the language but saved by the compiler: Robust safety via semantic back-translation,” Proceedings of the ACM on Programming Languages, vol. 10, pp. 1153–1182, 2026, issn: 2475-1421. DOI: 10.1145/3776682.
- S. Spies, N. Mück, H. Zeng, M. Sammler, A. Lattuada, P. Müller, and D. Dreyer, “Destabilizing Iris,” Proceedings of the ACM on Programming Languages, vol. 9, no. PLDI, pp. 848–873, 2025, issn: 2475-1421. DOI: 10.1145/3729284.
- T. Dardinier, M. Sammler, G. Parthasarathy, A. J. Summers, and P. Müller, “Formal foundations for translational separation logic verifiers,” Proceedings of the ACM on Programming Languages, vol. 9, no. POPL, pp. 569–599, 2025, issn: 2475-1421. DOI: 10.1145/3704856.
- M. Sammler, S. Spies, Y. Song, E. D’Osualdo, R. Krebbers, D. Garg, and D. Dreyer, “DimSum: A decentralized approach to multi-language semantics and verification,” Proceedings of the ACM on Programming Languages, vol. 7, no. POPL, pp. 775–805, 2023, issn: 2475-1421. DOI: 10.1145/3571220.
- M. Sammler, R. Lepigre, R. Krebbers, K. Memarian, D. Dreyer, and D. Garg, “RefinedC: Automating the foundational verification of C code with refined ownership types,” in Proceedings of the 42nd ACM SIGPLAN International Conference on Programming Language Design and Implementation, virtual: Association for Computing Machinery, 2021, pp. 158–174. DOI: 10.1145/3453483.3454036.
2 Group Members
- Klaus Krassnitzer, PhD Student since 2025
- Alice Lee, PhD Student since 2025
- Alexander Loitzl, PhD Student since 2025
3 Research
Low-level systems software like operating systems or hypervisors form the heart of modern computer systems. Such software provides critical components that ensure the reliability and security of the overall system. However, this means that bugs and vulnerabilities in these programs can lead to catastrophic failures and security problems. As a consequence, detecting and preventing bugs in low-level systems software is crucial for providing a reliable basis for the modern computer-based world.
One important approach to eliminate such bugs and vulnerabilities is formal verification. Unlike testing, which can show the presence of bugs for certain inputs, formal verification proves the absence of whole classes of bugs and vulnerabilities for all inputs. Concretely, formal verification allows one to prove that the low-level program behaves as described by a high-level mathematical specification.
The Sammler group develops novel approaches and tools that enable the formal verification of real-world programs, with a focus on low-level software. To achieve this, the group builds realistic models of programming languages like C or assembly. Then, the group develops verification methodologies that enable the verification of real-world code against these models. The work of the group ranges from advancing the theoretical foundations of verification to building practical verification tools that provide automatic verification while also generating machine-checkable proofs of correctness.
3.1 Current Scientific Projects/Directions
- Combining automated and foundational verification for real-world programs
- Reasoning about multi-language programs
- Understanding and verifying real-world C idioms
- Specification languages for separation logic-based verifiers
- Modular program logic design
- Non-interference-preserving secure compilation
- Building the foundation for the next generation of foundational and automated verifiers
3.2 Research in Focus
Modern software is routinely assembled from a mixture of trusted and untrusted components: web browsers execute third-party code, applications rely on plugins, and security-critical systems often link against libraries written in unsafe languages such as C. This raises a fundamental question: how can we guarantee that untrusted code cannot corrupt the state of trusted components or trigger unsafe behavior? A standard answer is encapsulation, for instance through sandboxing or compilation to hardware-supported capability machines such as CHERI. Yet giving formal guarantees for such protection mechanisms is challenging, especially when the source language itself is unsafe.
Together with researchers at the Max-Planck-Institute for Software System, we made a major step in improving the state-of-the-art by showing how to obtain robust safety for an unsafe, C-like language. Robust safety means that a trusted component behaves safely even when linked with arbitrary untrusted code. Prior work had established such guarantees only in settings where the source language already provides built-in encapsulation, making robust safety a contextual property of source programs. That assumption breaks down for C-like languages, whose contexts are too powerful because they may themselves exhibit undefined behavior.
The key contribution of this work is to show that unsafe source programs can nevertheless be “saved by the compiler.” Concretely, we developed a new proof framework based on semantic back-translation, which reasons semantically about all possible low-level attacker behaviors instead of requiring them to be represented syntactically in the source language. This framework is used to prove the soundness of a hardening compiler from an unsafe C-like source language to a CHERI-like capability assembly language. Overall, this work provides a novel and realistic foundation for proving security properties of systems written in unsafe languages but protected by secure compilation.
This work was published in POPL 2026: N. Mück et al., Proceedings of the ACM on Programming Languages, vol. 10, pp. 1153–1182, 2026. DOI: 10.1145/3776682.