Summary The e-Media Research Lab at KU Leuven is offering a fully funded, 4-year PhD scholarship focused on developing real-time, privacy-preserving camera systems. The project involves deploying and optimizing face de-identification pipelines on resource-constrained embedded hardware (tinyML) to anonymize individuals directly on-device while preserving essential attributes like age, gender, expression, and gaze.
Fully Funded 4-Year PhD Scholarship in AI & tinyML at KU Leuven, Belgium
Key Information
| Feature | Details |
| Designation | PhD Researcher / Scholarship |
| Research Area | Embedded AI, tinyML, Deep Learning, Hardware-Aware Neural Architectures, Privacy-Preserving Vision |
| Location | Leuven, Belgium |
| Employment Type | Full-time (Voltijds) |
| Duration | Up to 4 years (with 1 year pre-doc supported for non-EEA applicants, subject to intermediate evaluations) |
| Application Deadline | August 10, 2026, at 23:59 CET |
| Starting Date | October 1, 2026 (negotiable) |
Research Area: Embedded AI & tinyML for Privacy-Preserving Computer Vision
Job Description
Digital imagery introduces major privacy risks amplified by deepfake misuse and regulations like GDPR. The selected PhD researcher will focus on bringing face-swapping algorithms to off-the-shelf embedded platforms.
Key Responsibilities:
- Deploy face de-identification pipelines directly on resource-constrained hardware.
- Optimize underlying AI models based on strict latency, memory, and power constraints.
- Develop hardware-aware neural architectures using hardware-in-the-loop optimization.
- Implement efficient inference strategies (model compression, quantization, pruning).
- Build working demonstrators that showcase real-time anonymization in practice.
Eligibility & Qualifications
Candidates must be highly motivated with a strong interest in optimizing AI models on low-power devices. Requirements include:
- Education: Master’s degree in Electrical Engineering (or equivalent).
- Academic Excellence: Must be ranked within the top 10% of their class in both BSc and MSc, with exceptional grades.
- Technical Skills: * Solid background in deep learning (model training, inference, and evaluation).
- Experience with embedded hardware or AI accelerators (e.g., NVIDIA Jetson, Axelera, Qualcomm, Rockchip, or ARM-based platforms).
- Proficiency in Python, C/C++, and frameworks like PyTorch or TensorFlow.
- Interest in hardware-aware AI (quantization, pruning, efficient neural architectures).
- Soft Skills: Strong communication skills, fluency in English, and excellent interpersonal skills to thrive in an international team.
How to Apply
Interested candidates must submit their application via the online application tool. The following documents are required:
- A comprehensive CV explicitly mentioning BSc and MSc average grades.
- A Motivation Letter (maximum 1 page).
- Complete Transcripts of Bachelor and Master Degrees (the most recent transcript is acceptable if the MSc is not yet finalized).
- English Proficiency Test results (IELTS, TOEFL, or similar), if available.
Contact Information: For more information regarding the project or role, you can contact Prof. Dr. ir. Jona Beysens directly via email at jona.beysens@kuleuven.be.
