Research Engineer in Artificial Intelligence for Experimentation on the Combination of AI and Formal Methods

Context
The widespread adoption of AI across many areas of society calls for the development of techniques to make AI safer and more trustworthy. AI-based systems deliver impressive performance, but their lack of transparency makes them difficult to trust in industrial applications, as they provide no formal guarantees. Conversely, formal methods excel at logical reasoning and can prove properties of programs or detect errors, but they often face scalability limitations.
To address this challenge,
Inria and MERCE
(Mitsubishi Electric R&D Centre Europe) have launched a joint project, called
Défi FRAIME
, to explore, on the one hand, how formal methods can provide guarantees for AI systems and, on the other hand, how AI can help make formal methods more efficient and easier for practitioners to use. The vision of the project is to
intertwine formal methods and AI
to efficiently design safe systems. The project is driven by industrial use cases in MERCE’s areas of expertise.
This position is part of the FRAIME project, which provides a
stimulating collaborative research environment
involving 5 Inria teams and 2 MERCE teams, and will offer the successful candidate the opportunity to interact with around twenty recognized researchers in formal methods, software engineering, and AI.
The position aims to ensure the monitoring, deployment, evaluation, and sharing of AI technologies within the project, while providing researchers with hands-on expertise and consolidating the knowledge gained through experimentation.
The successful candidate will have access to following computing facilities:
Ilaas:
https://www.ilaas.fr/
Albert:
https://ia.numerique.gouv.fr/outils-ia/albert-api/
Abaca: (Inria internal computing infrastructure)
The successful candidate will be based at the Inria Centre at the University of Rennes and will interact frequently with the MERCE teams (located nearby) and with the Inria teams in Nantes and Paris involved in Défi FRAIME.
Assignment
The position consists in identifying, deploying, and evaluating the AI technologies most relevant to the project’s experiments, particularly around LLMs and SLMs. The successful candidate will continuously monitor technological developments and set up the infrastructure required to support the research activities. They will assist researchers by providing hands-on expertise in AI technologies and contributing to the project’s technical choices. Finally, they will promote the sharing, consolidation, and dissemination of the knowledge gained, notably through documentation, experiment summaries, and presentations at plenary meetings.
Main activities
Identification of technologies relevant to the experiments.
Monitoring of technologies related to the project (LLMs, SLMs).
Deployment of infrastructure that can be used throughout the project.
Implementation of evaluation methods for the experiments.
Providing researchers involved in the project with hands-on experience in AI technologies; proactively recommending technologies to be used.
Contributing to the sharing of know-how and knowledge on AI technologies within the project.
Consolidation and sharing of knowledge acquired within the project.
Production of summaries of the experiments conducted, and presentation at plenary meetings.
Skills
Required Technical Skills
Proficiency in RL / LLM / SLM technologies: fine-tuning (PEFT/LoRA), RAG, prompt engineering, quantization, inference optimization, technical infrastructure (local and remote deployment).
Ability to conduct experiments (benchmark design, reproducibility, and experiment tracking), analyze results, and evaluate AI systems.
Proficiency with collaborative development tools and practices (Git, continuous integration, software project management).
Master’s degree or equivalent in Computer Science, Artificial Intelligence, Applied Mathematics, or a related field.
Additional Assets / Desirable Extra Skills
Experience with formal methods would be a plus.
Knowledge of energy measurement technologies would be an asset.
Knowledge of agentic frameworks would be a plus.
Soft Skills
Professional proficiency in written and spoken English.
Strong written and oral communication skills.
Ability to work effectively as part of a team in a collaborative research environment.
Ability to work independently in project execution and problem solving.
Coordination skills.
Ability to interact with people from a variety of disciplines (AI, formal methods, software engineering).