Research Engineer in AI

The Inria Saclay-Île-de-France Research Centre was established in 2008. It has developed as part of the Saclay site in partnership with Paris-Saclay University and with the Institut Polytechnique de Paris .

The centre has 40 project teams , 32 of which operate jointly with Paris-Saclay University and the Institut Polytechnique de Paris; Its activities occupy over 600 people, scientists and research and innovation support staff, including 44 different nationalities.

The engineer will be part of the OPIS project team, reporting to team leader
Emilie Chouzenoux.

The project will be carried out in collaboration with E. Chouzenoux and J.-C. Pesquet (OPIS), and C. Lefort, CNRS research scientist at XLIM, Limoges. 

CARS microscopy (Coherent Anti-Stokes Raman Scattering) is an advanced nonlinear optical imaging technique that provides vibrational information about biomedical samples. The key advantage of CARS microscopy is its ability to deliver label-free spectroscopic information. The use of broad-spectrum laser sources, known as "supercontinuum," has been introduced to explore the full range of sample vibrations. This is referred to as M-CARS, or Multiplexed CARS. Additionally, the spectral detection capability of the dedicated instrument enables hyperspectral M-CARS imaging, allowing the collection and analysis of a wide light spectrum for each pixel in an image. Each pixel, therefore, contains a broad spectrum (1024 points), providing insights into the composition of substances in the scene.

Through this innovative hyperspectral M-CARS technique, we recently demonstrated that the "silent zone" of an M-CARS spectrum actually provides discriminative information about the sample—an area yet unexplored in the biomedical field. However, the current data processing method is cumbersome and prone to several challenges in data exploitation.

The hyperspectral images produced by the M-CARS solution generate a large volume of data due to the number of pixels in an image (a minimum of 10 × 10 pixels, typically 50 × 50, and up to 1024 × 1024), with each pixel containing 1024 spectral points.

The recruited engineer will be tasked with investigating artificial intelligence (AI) solutions to efficiently process these data and extract the discriminative information they contain. A hyperspectral M-CARS database recorded from muscle tissue will be used, where the myosin network is clearly identified by SHG, a well-known contrast method.

Between the myosin striations are actin striations, a protein that shows no SHG or vibrational signature. However, a discriminative signature has been observed in the "silent zone" of the M-CARS spectra, and intensity-level discrimination has been highlighted. The goal is to identify the spectroscopic signature using hyperspectral M-CARS and AI solutions. The mission involves developing an AI strategy to retrieve this discriminative information.

Subsequently, the task will extend to testing other biomedical samples, such as neurons, plants, or bacteria, to discriminate between different populations using the implemented AI method.

• Understand the image processing problem
• Analyze the database
• Deploy a supervised AI approach to solve the problem
• Write scientific reports
• Participate in scientific meetings with collaborators

• Proficiency in the Python programming language and the PyTorch or TensorFlow environment is required.
• Experience in machine learning / neural networks is strongly recommended.

• Canteen and cafeteria;
• Sports equipment;
• Transport reimbursement

Regarding professional experience