PHD in Multimodal Artificial Intelligence and On-Edge Deployment

Description

The “Integrated and Wireless Systems” Business Unit, based in Neuchâtel, Switzerland is looking for a PHD student.

Your mission is to develop next-generation AI methods that integrate diverse data modalities while remaining deployable on resource-constrained hardware platforms. The successful candidate will work on novel multimodal fusion methods, cross-attention architectures, and robust representation learning, with a strong emphasis on hardware awareness, efficient deployment, and model generalization. The project explores how imaging, sensor signals, structured clinical or contextual data, and temporal information can be combined into unified, interpretable, and transferable AI systems. Beyond algorithmic development, the research will investigate how these models behave under domain shifts and how they can be distilled into compact, deployable architectures suitable for edge devices.

Responsibilities

• Develop algorithms for integrating heterogeneous data types (e.g., imaging, signals, text, structured metadata).
• Design cross-attention, transformer-based, or graph-based models to capture cross-modal dependencies.
• Build spatio-temporal models for data streams with both structural and temporal dimensions.
• Create distillation pipelines where large multimodal models supervise smaller, hardware-efficient models.
• Investigate cross-modal distillation strategies that preserve multimodal semantics in compressed models.
• Innovation and Research: conduct in-depth research to advance the state-of-the-art in embedded training, exploring novel approaches and solutions.
• Industry Technology Transfer: Work on projects with the primary goal of transferring developed technologies and innovations to the industry, bridging the gap between academia and practical applications.
• Publication: disseminate research findings by publishing in top-tier academic journals and presenting work at prestigious conferences, contributing to the field's knowledge and recognition.
• Support proposal writing.
• Supervision of master students.

Qualifications

• Master’s degree in computer science, Electrical Engineering, Machine Learning, or a related field.
• Strong background in machine learning algorithms and deep learning.
• Experience with attention (cross) mechanism
• Proficiency in programming languages such as Python, C++, and frameworks like TensorFlow or PyTorch.
• Research Aptitude: demonstrated research experience through coursework or prior projects and show eagerness to engage in research activities and a willingness to learn and contribute to ongoing projects.
• Academic Excellence: strong academic record, with a history of outstanding performance in relevant coursework.

Interpersonal skills

• Natural curiosity, adaptability, collaborative approach.
• Autonomous, innovative thinker, problem-solving abilities, results oriented.
• Strong organisation and communication skills.