PhD in Embedded Software for Dynamic Neural Networks
PhD in Embedded Software for Dynamic Neural Networks at KU Leuven
The Embedded Systems unit of EAVISE at KU Leuven is recruiting a highly motivated PhD researcher to work on a novel research direction in Embedded AI: enabling neural networks to dynamically optimize themselves after deployment. The project addresses a major limitation of today’s edge and embedded AI systems, where models are optimized before deployment but remain static during their lifetime. This can leave energy consumption, memory usage, and compute requirements fixed even as conditions change. The proposed research aims to build software methods that let deployed deep neural networks progressively reduce their resource footprint while maintaining application performance.
The work sits at the intersection of embedded software, compiler systems, AI deployment, and hardware-software co-design. Research themes include runtime-reconfigurable neural network kernels, performance models for adaptive AI workloads, on-device search and optimization engines for memory-aware deployment, compiler and runtime support for dynamic neural network topologies, and energy-efficient embedded AI frameworks for RISC-V multicore and heterogeneous computing platforms. The project combines fundamental research with practical validation on modern embedded AI hardware.
EAVISE is a multidisciplinary research group at KU Leuven active in AI, embedded computing, computer vision, and edge intelligence for real-world applications. The Embedded Systems unit focuses on energy-efficient computing, AI deployment tools, and co-design methods for future intelligent devices such as wearables, smart sensors, hearables, and IoT nodes. The successful candidate will join a strong research environment with opportunities for doctoral training, international collaboration, and conference participation.
Eligibility highlights:
applicants should have a master’s degree in electrical engineering, computer engineering, or a related equivalent field. Master’s students nearing graduation are also welcome to apply. Strong C/C++ programming skills in embedded contexts, knowledge of compilers or software engineering, and familiarity with computer architectures and embedded platforms are strongly preferred. Background in deep learning and AI is a plus. Excellent English and good communication skills are required, along with the ability to work in an international team.
Funding:
the position is a PhD scholarship for up to 4 years, subject to positive intermediate evaluations. KU Leuven indicates a competitive salary or tax-free PhD grant, including reimbursement of commute costs.
Application:
candidates must apply through the official application channel. In addition to the standard documents, the application should include a maximum 4-page summary of the master’s thesis (or ongoing thesis work for those not yet graduated), a maximum 3-page summary of relevant technical projects, a CV showing study duration and expected graduation date if applicable, and up to three references.
For more information, interested candidates may contact Prof. dr. ir. Manuele Rusci at [email protected].
