Raphael d'Assignies
Raphael d'Assignies is a lawyer, developer and specialist in digital trust systems. His career stands out for more than twenty-five years at the interface of law, computing and cybersecurity. He headed CELOG, dedicated to signature, timestamping and secure preservation architectures, chaired the Agence pour la Protection des Programmes and served as publication director of Expertises des systèmes d'information and Legalis.net.
Of counsel at Ydès since 2023, he advises on digital issues, intellectual property, security and artificial intelligence. He also teaches AI applied to law at HEAD and, since 2026, at the University of Tours. His courses combine how models work, risks, ethics, regulation, prompt design and the delivery of professional use cases.
He created Legal GPT in 2022 and OpenLegi in 2025. OpenLegi facilitates automated access to official sources such as Légifrance, the French National Business Register, EUR-Lex, BODACC and BOFiP. Raphael also publishes open-source components, including the Python library pylegifrance and an MCP server for Légifrance, to make legal data usable by specialised assistants and agents.
For Lelink, he brings the ability to move from concept to prototype to product. He designs assistants and agents tailored to law firms, structures access to sources, secures data flows and translates confidentiality, traceability and reliability requirements into concrete architecture choices.
- —IAPP, since 2024: training track in artificial intelligence governance
- —École de Formation du Barreau, 1997-1998: CAPA and professional training as a lawyer
- —IHEDN, 2015: training session on defence and security issues
- —Conservatoire national des arts et métiers, 2014-2018: training in computer science
- —Université Paris-Sud, Faculté Jean-Monnet, 1991-1998: master's degree in law
- —Université Paris-Panthéon-Assas, 1995-1997: master's degree in law
- —Technical certifications: Machine Learning (Stanford Online, 2021); Mathematics for Machine Learning — linear algebra and multivariate calculus (Coursera, 2020); Introduction to Python for Cybersecurity (Infosec, 2024)
Creation of a platform and a technical access layer to official legal sources, allowing AI tools to consult up-to-date, traceable and structured data rather than depending on an opaque corpus.
Publication of open-source building blocks that simplify querying the PISTE/Légifrance API and its use by assistants or agents — hands-on expertise in integrating public data, Python and AI.
Development of legal use cases since 2022 and teaching for lawyers, in-house counsel and students, with constant attention to hallucinations, evaluation, confidentiality and human supervision.