Chief Scientist, Professor, Google Chair
I’m Maxime Cordy, Chief Scientist, Professor, and Google Chair in Trustworthy and Sovereign Artificial Intelligence working at SnT - The Interdisciplinary Centre for Security, Reliability and Trust, University of Luxembourg. I lead impact-driven research at the intersection of trustworthy AI, software engineering and industry applications. I am your trusted partner to turn ambitious ideas into dependable real-world systems.
As deputy head of the Security, Reasoning and Validation group, I lead a team working on AI quality, security and innovative applications. I collaborate with companies and public institutions across finance, energy, cybersecurity, manufacturing, official statistics and space—from fundamental research to deployed prototypes and spin-offs.
From meaningful research to collaborative delivery and tangible adoption.
We conduct rigorous research with a real-world perspective and a clear ambition: creating knowledge, methods and tools that deliver tangible impact for industry and society.
Through SnT, we build long-term collaborative projects with companies and public institutions, combining their domain expertise with our research capabilities to solve strategic challenges.
We turn promising research into reusable open-source tools, mature technologies and new ventures that can be adopted beyond the laboratory.
I advise organisations on innovation and technology, from framing opportunities and technical choices to shaping prototypes and actionable roadmaps.
A selection of academic, industrial and technology-transfer initiatives.
Developing a security sandbox for agentic AI applied to cybersecurity use cases.
Developing a sovereign software stack to make AI agents tested, reliable and secure.
Leading SnT’s R&D and the AI Testing Sandbox: an open-source platform for organisations to design secure, on-premises environments and assess AI safety, performance and regulatory compliance.
Developing LLM agents for complex tabular data, benchmarking causal-discovery libraries and building retrieval-augmented systems for financial market intelligence.
Developing methods and tools for certifying trustworthy data-intensive and AI systems.
Automated monitoring and continuous improvement of AI-driven financial decision-making systems, with transparency, robustness and compliance at the centre.
Developing assessment and improvement methods for Vision-Language Models applied to autonomous driving.
Forming an interdisciplinary research team and developing a software platform to assess the compliance of AI systems used in financial services.
Automating the evaluation of complex system designs using artificial intelligence and formal verification techniques.
Developing realistic threat models and reliable methods for assessing the adversarial robustness of tabular machine-learning systems.
Building an optimisation framework for the design and evaluation of complex space-program models.
Exploring high-value public-sector use cases and developing machine-learning technologies for trustworthy, modern official statistics.
Applying machine learning and operational data to improve predictive maintenance and automation in manufacturing plants.
Developing efficient methods to sample representative system configurations from very large and complex configuration spaces.
Improving how real-world machine-learning systems generalise as data and operating conditions evolve.
Applying machine learning to strengthen the reliability, security and operation of smart-grid infrastructures.
Developing an open-source MLOps platform for building and operating machine-learning systems.
Advancing semi-symbolic model checking to make verification practical for variability-intensive software systems.
Developing systems and methods that protect valuable digital assets against extraction and misuse by malicious neural networks.
Optimising electric-vehicle charging schedules using customer flexibility, operational constraints and incentive mechanisms.
Searching for misleading inputs and evaluating adversarial machine-learning systems under realistic domain-specific constraints.
Creating model-based methods to optimise the configuration of highly variable software-intensive systems.
Combining machine learning, epidemiological models and search-based methods to support evidence-informed pandemic exit strategies.
Studying adversarial machine learning and practical robustness for AI-supported financial decision-making systems.
Developing techniques for the systematic testing, quality assurance and security evaluation of self-learning systems.
Co-founded a company transferring research in AI and configurable systems into software for customised decision-making, recommendation and optimisation. The company continues today with 14 employees.
Stories and insights from research, collaborative projects and the journey from scientific ideas to real-world applications.
I advise PhD students and postdoctoral researchers across trustworthy AI, AI security, software engineering and industry-led research. Several former students received distinguished thesis recognition and now contribute across academia and industry.
Olivier Zeyen · Now working at ILNAS
Salijona Dyrmishi · excellent thesis award; Senior AI Security Researcher at SWIFT
Thibault Simonetto · Data engineer at POST.lu
Fabien Bernier · Research Associate at SnT
Qiang Hu · Associate Professor at Tianjin University
Salah Ghamizi · excellent thesis award; Principal Investigator at LIH and Research Scientist at SnT
Martin Gubri · excellent thesis nominee; Assistant Professor at École Polytechnique
Karim Tit · Yu-Lin Huang · Tagir Fabarisov · Joel Ky · Xiongfei Wu · Hamid Bostani · Tania Carvalho · Ali Kordia · Safaa Moallim · Fabien Bernier · Olivier Zeyen · Lorenzo Cazzaro · Firas Bayram · Simon Thrane Hansen · Sami Lazreg · Salijona Dyrmishi · Thibault Simonetto · Zeming Dong · Laura Bussi
PhD student
Co-advised Industrial PhD student
PhD student
PhD student
Industrial PhD student
Industrial PhD student
Industrial PhD student
PhD student · FNR AFR grantee
PhD student
PhD student · FNR AFR grantee
Industrial PhD student
Industrial PhD student
Co-advised PhD student
Alongside research, I shape curricula, research communities and institutional initiatives that connect AI, cybersecurity and society.
Study Programme Director for the University of Luxembourg’s professional master programme since 2025.
Creator and main lecturer for the AI & Cybersecurity course offered to students in the Erasmus Mundus Joint Master in Cybersecurity and the Master in Cybersecurity and Cyber Defence.
Leadership-team member of the Luxembourg FutureFinTech research centre and external advisor to the National AI Strategy.
University representative on the EuADS board and main University of Luxembourg contact for The Guild’s AI/Digital Research Heads Working Group.
Video and written conversations about trustworthy AI, cybersecurity, energy and research with real-world impact.
How researchers contribute to AI policy while building the technical tools that companies will use to assess their systems.
Read feature ↗Presenting Luxembourg research and technology designed to strengthen the reliability and trustworthiness of AI systems.
Read feature ↗How SnT is advancing sovereign, trustworthy artificial intelligence through research, infrastructure and collaboration.
Read interview ↗Exploring how language models can evolve beyond writing to work reliably with complex financial and tabular data.
Read interview ↗Exploring how research and industry combine their capabilities to strengthen space-weather intelligence.
Read feature ↗A research partnership using machine learning to improve forecasts and protect critical infrastructure.
Read interview ↗How five years of AI-security research developed into technology designed for real-world protection.
Read feature ↗An in-depth conversation about SVALINN, adversarial AI and protecting valuable digital assets.
Read interview ↗Discussing the INDUCTIVE project, applied research and how academia can support companies and society.
Watch & read ↗Public-facing coverage of research into more flexible, efficient electric-vehicle charging.
View on RTL ↗Coverage of our research on generating realistic adversarial examples that respect the constraints and semantics of tabular data.
Read feature ↗