Thomas Dietzen

This PhD position at KU Leuven, within the EAVISE research group, focuses on developing domain-aware audio representations for low-resource settings. EAVISE is a multidisciplinary group operating at the intersection of artificial intelligence, embedded systems, computer vision, and sound processing, and is part of the Electrical Engineering Department (ESAT) and Computer Science Department at KU Leuven. The project aims to advance audio representation learning by designing model structures and learning objectives that explicitly incorporate domain knowledge, particularly for scenarios where training data is limited and compact models are necessary. Application areas include predictive maintenance, process monitoring, environmental monitoring, and safety-critical event detection. The research will address challenges in data efficiency, computational efficiency, and domain generalization by disentangling sound representations based on signal characteristics. The successful candidate will join an international research environment with access to state-of-the-art infrastructure and benefit from close mentorship, participation in courses, workshops, and conferences, and opportunities for collaboration with academic and industrial partners. The position offers a competitive salary or tax-free PhD grant, with an initial one-year appointment and potential extension up to four years. Candidates are expected to conduct research, contribute to project planning, disseminate findings, assist in supervising Master’s students, and participate in limited teaching activities. Applicants must have a Master’s degree in electrical or computer engineering (or a related field), strong mathematical background, coursework in signal and sound processing, proficiency in Python, and excellent English communication skills. Research experience in sound processing and additional programming skills are advantageous. KU Leuven is committed to diversity and inclusion, providing a supportive and respectful environment for all researchers. The application deadline is December 3, 2025, and applications must be submitted via the online tool.

This PhD position at KU Leuven focuses on domain-aware audio representation learning for low-resource settings, embedded within the multidisciplinary EAVISE research group. EAVISE operates at the intersection of artificial intelligence, embedded systems, computer vision, and sound processing, and is part of the Department of Electrical Engineering (ESAT) and the Department of Computer Science. The group offers a stimulating international research environment, access to state-of-the-art infrastructure, and strong industry and academic connections. The project aims to advance audio representation learning by designing model structures and learning objectives that explicitly incorporate domain knowledge, particularly for scenarios where training data is limited and model compactness is essential. Application areas include predictive maintenance, process monitoring, environmental monitoring, and safety-critical event detection. The research will focus on disentangling sound representations based on signal characteristics, improving data and computational efficiency, and enhancing domain generalization. The successful candidate will conduct research on audio representation learning, participate in project planning and meetings, collaborate with academic and industrial partners, disseminate findings through publications and conferences, and contribute to follow-up research proposals. Additional responsibilities include assisting with Master’s thesis supervision, performing limited teaching activities, and enrolling in the Arenberg Doctoral School’s doctoral training program. Applicants should hold a Master’s degree in electrical or computer engineering (or a closely related field), with strong mathematical skills and coursework in digital signal processing, sound processing, and machine learning. Research experience in sound processing is highly valued, and candidates with above-average grades, awards, or publications are encouraged to apply. Proficiency in Python is required, with MATLAB or C/C++ experience considered a plus. Excellent English proficiency and strong communication skills are mandatory. The position offers a competitive salary or tax-free PhD grant, an initial one-year appointment renewable up to four years, flexible working hours, and opportunities for remote work. The application deadline is December 3, 2025. Applications must be submitted via the KU Leuven online application tool and should include a cover letter, CV, transcripts, and a link to the master’s thesis if available. For further information, contact Prof. dr. Thomas Dietzen at [email protected]. KU Leuven is committed to diversity, inclusion, and equal opportunity, fostering a respectful and socially safe environment for all researchers.

The EAVISE research group at KU Leuven, located on Campus De Nayer Sint-Katelijne-Waver, offers a multidisciplinary environment within the divisions of Processing Speech and Images (PSI) in the Electrical Engineering Department (ESAT) and Declarative Languages and Artificial Intelligence (DTAI) in the Computer Science Department. EAVISE specializes in demand-driven applications of advanced algorithms for artificial intelligence, embedded systems, computer vision, and sound processing, with strong ties to industry and a robust research infrastructure. This PhD position focuses on modeling and processing audio signals under motion, a challenging area where time-varying delays (known as warping) affect the time axis of audio signals. While stationary audio processing has advanced significantly, motion-induced warping—responsible for effects like the Doppler shift—is rarely explicitly modeled. The project aims to develop novel models and processing approaches for audio signals affected by warping, covering both analysis and synthesis tasks. Methodologies include stochastic signal processing and machine learning, with hybrid physics-guided and data-driven frameworks. Research directions may involve modeling warping interactions with signal transformations, estimating motion parameters from audio, moving-microphone measurements of room acoustics, and efficient synthesis and auralization of dynamic acoustic scenes. The successful candidate will join the international EAVISE group and benefit from a stimulating research environment, access to state-of-the-art infrastructure, and opportunities for collaboration with academic and industrial partners. The standard PhD duration at KU Leuven is four years, with a starting date as soon as practical, preferably in summer 2026. The position offers a competitive salary or tax-free PhD grant, flexible working hours, and the possibility of remote work. The initial appointment is for one year, with extensions possible up to four years. Applicants must hold a Master’s degree in electrical or computer engineering (or a closely related field), possess a solid foundation in mathematics, and have coursework in digital signal processing, sound processing, and machine learning. Research experience in sound processing is highly valued, especially for candidates with above-average grades, awards, or scientific publications. Proficiency in Python is required; MATLAB or C/C++ experience is a plus. Excellent English language proficiency and strong communication skills are essential. The candidate will conduct research, participate in meetings, collaborate with partners, disseminate findings, supervise Master’s students, perform limited teaching, and enroll in the Arenberg Doctoral School. KU Leuven is committed to diversity, inclusion, and equal opportunity. Applications must be submitted online by April 13, 2026. Required documents include a cover letter, CV with GPA and relevant courses, transcripts, and a link to the Master’s thesis if available. For further information, contact Prof. dr. Thomas Dietzen at [email protected].

The EAVISE research group at KU Leuven, located on Campus De Nayer Sint-Katelijne-Waver, offers a multidisciplinary environment within the divisions PSI (Processing of Speech and Images) of the Electrical Engineering Department (ESAT) and DTAI (Declarative Languages and Artificial Intelligence) of the Computer Science Department. EAVISE specializes in demand-driven applications of artificial intelligence, embedded systems, computer vision, and sound processing, leveraging a robust research infrastructure and international collaborations. This PhD position focuses on knowledge-guided audio representation learning for low-resource settings. Recent advances in machine learning for audio have prioritized learning from raw data using large-scale models, but these approaches are often data-hungry and computationally intensive. Many real-world applications, such as predictive maintenance, process monitoring, and environmental monitoring, require models that generalize well with limited training data and operate efficiently under deployment constraints. The project aims to adapt self-supervised audio representation learning to low-resource environments by designing model structures and learning objectives that explicitly incorporate prior knowledge about acoustic signal structure. By making the acoustic structure explicit in both model design and training, the research seeks to improve data efficiency, computational efficiency, robustness, and domain generalization of learned audio representations. The successful candidate will join the international EAVISE group and benefit from a stimulating research environment, access to state-of-the-art infrastructure, and opportunities for collaboration with academic and industrial partners. The standard PhD duration at KU Leuven is four years, with a starting date as soon as practical, preferably by September 2026. Applicants must hold a Master’s degree in electrical or computer engineering (or a closely related field), possess a strong foundation in mathematics (including matrix algebra), and have completed coursework in digital signal processing, sound processing, and machine learning. Research experience in sound processing is highly valued, especially if accompanied by above-average grades, awards, or scientific publications. Proficiency in Python is required, and experience with MATLAB or C/C++ is advantageous. Excellent English language proficiency and strong communication skills are essential. The position includes conducting research, project planning, participation in meetings, collaboration with partners, dissemination of findings through publications and conferences, proposal development, supervision of Master’s students, limited teaching activities (up to two hours per week), and enrollment in the Arenberg Doctoral School’s doctoral training program. The offer includes a PhD degree from KU Leuven after successful completion, comprehensive scientific education, close mentorship, participation in courses and conferences, a competitive salary or tax-free PhD grant (depending on funding scheme), an initial one-year appointment with possible extensions, flexible working hours, and remote work options. Applications must be submitted via the KU Leuven online application tool by May 12, 2026. Required documents include a cover letter, CV (with GPA, coursework, thesis link, and publications), transcripts, and language certificates. If the Master’s degree is pending, indicate the expected graduation date. For further information, contact Prof. dr. Thomas Dietzen at [email protected]. KU Leuven is committed to diversity, inclusion, and equal opportunity, fostering a respectful and socially safe environment for all.