Postdoc Position - Deep Learning on 4D Data

University of Bordeaux
Bordeaux, France
Job Type: 
Full Time
Closing Date: 
Friday, September 1, 2017

Ginesislab is seeking a postdoctoral researcher to carry out research on the extraction of phenotypes from brain imaging data (structural, diffusion and functional MRI). He or she will have to develop and validate these approaches using different cohorts (BIL & GIN, i-Share, UK-BIOBANK, Human Connectome Project). In a first phase, he or she will be responsible for overseeing the analysis and performing validation tests to characterize the variance and reproducibility of the phenotypes calculated for each subject and each cohort. In a second phase, he or she will work on methods for predicting the anatomo-functional organization of an individual by using classification methods, in particular Deep Learning, using phenotypes identified during the first phase. In addition to publishing this work in leading international journals, the main objective is technology transfer to the company.

Who is Ginesislab? Created under the aegis of a Joint Public-Private Laboratory program and supported by the CEA, Ginesislab associates an academic laboratory, the Groupe d'Imagerie Neurofonctionnelle (GIN) specialized in neurosciences, and Cadesis, specialized in the deployment of information systems. Ginesislab's main lines of research are: 1) integrated management of brain imaging data processing, design and automation of local calculations or remote computing centers; (2) the development and integration of statistical methods for the analysis of intra- and inter-cohort variability in order to classify and detect the healthy and pathological determinants of the anatomo-functional organization of the human brain.

Requested knowledge. The candidate presenting a curriculum dealing with "computer sciences data" or "neurosciences" will be evaluated according to the following criteria:

  • Competence in medical imaging and processing of anatomical and functional images;
  • Competence in applied mathematics / statistics, multivariate analysis, classification and prediction;
  • Knowledge of development methodology and programming languages (knowledge of Matlab, Python and their environments will be appreciated);
  • Rigor, autonomy and curiosity to get involved in a multidisciplinary and multi-team project;
  • Good communication skills in English (oral and written).

The application must include:

  • A complete CV integrating the candidate's experience as well as his/her academic training;
  • Letters of recommendation from people who can support the candidate;
  • A detailed and personalized letter of motivation for the position;
  • The candidate's publications (PDF format).

Candidates will be evaluated by analysis of their application and an interview by the co-directors of Ginesislab.

Starting date: About September 2017
Duration: 2.5 years (annualized contract)
Location: Bordeaux, France
Salary: About 2000€ net/month depending on experience

Please send your application to:  Marc Joliot