Missions

The Artificial Intelligence Engineer will be responsible for developing innovative solutions for processing and analysing complex data. The main objective is to install a MLOps infrastructure at LPC Caen, to develop machine learning models on nuclear physics data, in particular auto-encoders for neutron/gamma discrimination, and to optimise these models for execution on embedded systems.

Activités

- Integration of MLOps tools at LPC Caen
- Development of an auto-encoder for neutron/gamma discrimination
- Identification of charged particles (PSA) using deep learning
• Development of an auto-encoder for data flow reduction and data quality monitoring
• Size and performance model optimisations for embedded system execution
- Documenting, maintaining and deploying the methods and tools developed (GIT, CI/CD)
- Present results in workshops and international scientific conferences
- Organising training courses in English for laboratory staff to ensure knowledge and skills the transfer

Compétences

- Proficient with MLOps, MLflow and DVC tools
- Proficient in programming with languages such as Python and C/C++.
- Knowledge of AI frameworks and libraries such as TensorFlow, PyTorch, Keras, and ONNX
- Experience with machine learning and deep learning algorithms
- Knowledge of model optimisation techniques for embedded systems (quantisation, pruning, etc.)
- Experience of deploying AI models on embedded platforms (e.g. NVIDIA Jetson, Raspberry Pi, etc.)
- Use of collaborative development tools such as Git.
- Ability to manage development projects and collaborate effectively in a team.
- Fluency in English
- A good knowledge of nuclear physics and associated detection is an asset.

Contexte de travail

The research engineer will be assigned to the Caen Corpuscular Physics Laboratory, a joint research unit with a staff of 108. The laboratory's main activity concerns fundamental research: studying the properties of the forces at work in atomic nuclei in order to understand the mechanisms responsible for their formation. The LPC-Caen is made up of 7 research groups and 5 technical departments. The research engineer will work in the software department, which has 7 employees. He/she will participate in the development of the SINAPSE project (SImulateur Numérique pour l'Apprentissage Profond sur Système Embarqué). This project is a simulator that will enable experimental data analysis concepts to be validated using artificial intelligence. The aim is to use neural networks to improve the identification of particles from nuclear reactions, auto-encoders to compress data and detect anomalies in real time, and to develop software tools to present the results at conferences and seminars.

Contraintes et risques

This project does not present any particular professional risks.
It may be necessary to travel to present the results.