Vous êtes ici

R&D scientist in Multimedia AI for Social Media Analytics


The LINKMEDIA team at IRISA , public research laboratory, works on the development of technology for the description of and access to multimedia data leveraging content analysis: computer vision, speech and language processing, audio processing, fusion, data mining and AI. Examples of research activities in the team include (but are not limited to) named entity detection in texts, speaker segmentation, similar image search, object detection in images, multimodal and cross-modal content retrieval. We also study how these foundational functionalities can be assembled to organize multimedia collections and develop prototype applications facilitating analytics and knowledge acquisition from such collections.

LINKMEDIA is involved in the international EIT Digital project CREEP 2 (Cyberbullying Effects Prevention ) involving research labs and private companies to develop innovative technology facilitating the detection, prevention and management of cyber harassment between teenagers on social networks. In the context of this project, we are hiring as soon as possible a specialist in multimedia content analysis (text and/or image and/or multimodality) in charge of R&D activities on image analysis and multimodal fusion within the project.


As a member of the LINKMEDIA team and in close collaboration with CREEP’s international partners, you will be in charge of designing, implementing and benchmarking multimodal approaches to detect and categorize situations of cyber harassment, exploiting images posted on social networks along with the corresponding text. We will in particular rely on deep learning techniques developed in the team for the fusion of heterogeneous modalities (V. Vukotić et al. A Crossmodal Approach to Multimodal Fusion in Video Hyperlinking. IEEE Multimedia, 2018). R&D activities will primarily focus on the design of neural fusion architectures, for instance combining early/late fusion, and allowing to justify the decisions taken (explainability). We will in particular investigate the use of attention mechanisms to do so.


Profil / compétences: 

Applicants should ideally hold a PhD or a significant experience in R&D projects involving deep learning and/or multimedia content analysis. Knowledge of the French language is not required but is however a bonus Applications from newly graduated Master students or equivalent will also be considered, provided they have significant qualified training in deep learning.

Type of contract: temporary position starting a.s.a.p. up to Dec. 31, 2019



Lieu de travail: 
IRISA Rennes
Type de contrat: 
Durée du contrat (en mois): 
Salaire Brut / Mens €: 
between 25 k€ and 32k€ gross salary yearly depending on experience, including social benefits
Date prévisionnelle d'embauche: 
As soon as possible

To apply, please send resume, motivation letter and contact information of personalities able to recommend your application to Ewa Kijak : ewa.kijak[at]irisa.fr and Guillaume Gravier : guillaume.gravier[at]irisa.fr with the subject “Social Media Analytics Application”