Missions

As part of this position, the successful candidate will design and develop a self-supervised learning model for a new 3D super-resolution optical microscopy system.
This interdisciplinary work combines machine learning, statistics, optics, electronics, image processing, chemistry, and biology. The recruited engineer will lead one aspect of the system design while working in close synergy with the rest of the team on the other components.
The project will initially focus on a first microscope prototype as a proof of concept. The next objective will be to extend the method toward the development of a foundation model for 3D signal decoding, enabling super-resolution imaging of new biological samples at depths beyond the current state of the art.

Activities

The successful candidate will contribute to various aspects of the project, depending on their background and expertise, with a stronger involvement in one of the activities.
Throughout the project, the main tasks will include:
-Assessing the theoretical performance of the system through modeling,
-Developing and training a self-supervised learning model,
-Evaluating the model performance using both simulations and experimental data,
-Extending the approach from a task-specific model toward a foundation model,
-Benchmarking the results against state-of-the-art methods,
-Preparing progress reports and/or scientific publications,
-Presenting the results at national and international conferences.

Skills

Technical skills are expected in machine learning, applied mathematics, or image processing, combined with strong programming expertise in Python.
The recruited candidate should ideally have prior experience in machine learning, and in particular in self-supervised learning. Experience or familiarity with unlabeled data problems, pretext tasks, and pre-training, transfer learning, and fine-tuning approaches will be highly valued.
Beyond technical skills, the candidate is expected to show a genuine interest in the scientific implications of machine learning. They should be able to communicate regularly about their work and demonstrate a strong interest in teamwork and interdisciplinarity.

Work Context

This work will take place within the NanoBio team at ISMO (a joint CNRS / Université Paris-Saclay research unit) as part of the ERC project TimeNanolive.
The NanoBio team develops novel fluorescence microscopy modalities that push the limits of observation both in terms of acquisition speed and imaging depth, with applications ranging from biology to the study of nanomaterials.

These developments lie at the intersection of multiple disciplines, combining expertise in optics, electronics, image/data processing, chemistry, and biology.
With several European funding programs, the team is building a data science and machine learning group to foster innovation across the various microscopy-related fields.
The position is therefore particularly suited to candidates eager to work at the heart of this interdisciplinarity, with training in one of the major domains listed above and a desire to explore the others.
The laboratory is equipped with several single-molecule localization microscopes and has access to shared facilities, including biosafety level 1 and 2 cell culture labs and mechanical/electronics workshops, to support the completion of the project.
The contract duration can be extended.

Constraints and risks

laser beam
cell culture