Presentation of the project : Financial fraud generates major costs and evolves constantly with digitalisation and new payment channels.
This project develops interpretable and operationally relevant methods to detect fraud in insurance, social contributions and banking.
The research uses large, heterogeneous and often sensitive datasets to identify anomalous behaviour, suspicious networks and emerging fraud patterns while controlling false positives and preserving model interpretability.
Work will cover modelling under class imbalance, (semi-)supervised and cost-sensitive learning, network and graph methods to uncover coordinated schemes, and the development of explainable and robust models suited to regulated environments.
Particular attention will be paid to adaptive strategies as fraud techniques evolve and to temporal dynamics.
Presentation of the Service / laboratory : The postdoctoral researcher will join the Laboratoire d'Économie d'Orléans (LEO) and collaborate with academic and institutional partners.
The environment supports quality research, with access to high-performance computing, proprietary fraud datasets, and a multidisciplinary team.
The laboratory supports scientific publication and international conference participation.
Attention will be paid to data confidentiality, responsible use and research reproducibility.
Presentation of the missions : We seek a postdoctoral researcher to develop machine learning and deep learning models for fraud detection in insurance, social contributions and banking.
The work will include developing interpretable
ML/DL methods for anomaly detection, analysing and preprocessing large datasets from diverse financial sectors, conducting comparative studies of detection methodologies, and collaborating with industry partners to validate models on proprietary data.
The researcher will contribute to 2-3 publications in peer-reviewed international journals and present at major conferences.
Scientific autonomy will be significant, with mentoring from senior researchers and access to laboratory infrastructure.
Présentation of the profile : We are seeking a candidate who holds a Ph.D. in Data Science, Machine Learning, Econometrics, Finance, Statistics, or a related field, or who is expected to defend their doctoral thesis in the near future.
From a technical perspective, advanced expertise in machine learning and deep learning (ML/DL) is expected.
The candidate should be proficient in programming with Python, R, or equivalent tools, have experience working with large-scale datasets and feature engineering, and be familiar with explainable AI (XAI) methods.
Experience with network or graph analysis would be an asset.
From a professional perspective, the candidate should be able to conduct independent research, demonstrate excellent written and spoken English, work effectively in a collaborative environment, and handle confidential data appropriately.
A strong interest in fraud detection or financial risk analysis is important.
Previous experience working with insurance, banking, or social protection data, as well as proficiency in French, would be considered an asset.
Application file: - a CV including a list of publications;
- a cover letter outlining the candidate’s research interests and fit for the position (maximum 2 pages);
- one representative research paper.