Role Summary The Perception and Activity Understandin g group at Idiap seeks a highly motivated PhD student to join an innovative French–Swiss SNSF project (AIMANT) focused on developing advanced social-AI models for behavior understanding in the context of autism early screening, diagnosis, and monitoring. While significant progress has been made in activity and non-verbal behavior recognition, current methods still struggle with adult–child interactions captured in (close to) naturalistic conditions, such as those encountered in autism assessment sessions (ADOS, ESCS). These challenges arise from the complexity and subtlety of the target behaviors, including gesture- and attention-related cues (e.g., gaze, pointing, manipulation), infant micro-expressions, and motor stereotypes. This PhD project aims to develop next-generation multi-behavior coordination models (face and body gestures, gaze, audio, semantics) with temporal interpretation, at both the individual and interaction (child–clinician) levels and as a key objective to leverage these multimodal behavioral models to derive robust and clinically meaningful ASD profile scores from video data. The successful candidate will play a central role in addressing these challenges by contributing to model design, developing innovative datasets and benchmarks, and exploring cutting-edge machine learning strategies, including multi-task learning, unsupervised learning, and model distillation to unify behavioral representations across behaviors and heterogeneous datasets. Conditions: The position is fully funded for four years, contingent on satisfactory progress, and will lead to a PhD dissertation. The selected candidate will be enrolled as a doctoral student at EPFL, subject to acceptance by the Electrical Engineering Doctoral School (EDEE) ( http://phd.epfl.ch/applicants ). The anticipated start date is between June and September 2026. Salaries follow EPFL standards. Your Profile The ideal PhD candidate holds an MSc degree in computer science, engineering, physics, or applied mathematics. A strong foundation in statistics, linear algebra, signal processing, machine learning, and programming is essential, along with solid experience in deep learning. We are looking for candidates with strong analytical skills, excellent written and oral communication abilities, and a clear interest in working with medical behavioral data to address impactful real-world challenges. The ability to work in a collaborative, interdisciplinary environment with research and application partners is essential. How to apply : Please provide the following information, as two separate pdf files: File 1: (i) updated CV (ii) grades from your Bachelor and Master's degrees. If you do not yet have your Master’s diploma, please send a certificate issued by your current institution of higher education that confirms your enrollment in a Master’s program, as well as current grades. File 2: (i) a personal motivation letter (max 2 pages); (ii) a sample of published work (Master thesis, article), if available; (iii) reference letters (if available). Contact details of 3 reference persons. Application deadline : Screening of applications will begin on April 10th, 2026, and will continue until the position is filled. All queries related to the advertised position can be sent to Dr. Jean-Marc Odobez (jean-marc.odobez@epfl.ch). About Idiap Idiap is a research institute of national importance that engages in fundamental research, education and technology transfer in artificial intelligence. Idiap offers competitive salaries and conditions in a dynamic, multicultural environment. Idiap is located in Martigny, Valais, a scenic region in the south of Switzerland, surrounded by the highest mountains of Europe, and offering exceptional quality of life, exciting recreational activities, including hiking, climbing and skiing, as well as varied cultural activities. At Idiap, we believe that diversity fosters creativity and innovation. We are committed to equality of opportunity, to being fair and inclusive, and to being a place where we all belong. If you believe that you could excel in this role, but you are not sure you meet all the required criteria, we very much welcome your application. Required languages English - Level advanced Réf: c7b9639d-7567-4da4-a193-9955c6382352