In the sonar framework, developed here in Mouscron, we integrate software components from the antenna electronics side up to the generation of human interpretable images of the seafloor. The image generation feature uses complex mathematical algorithms which have been implemented in GPU code (Nvidia CUDA).
The software is tightly integrated with the framework; we usually test the implementation in the full context of execution the entire framework. We are currently working to make the GPU code architecture more modular. This could help with creating more focused tests for different stages of the GPU implementation.
For this internship, you will join the software development team working on the GPU code and help with restructuring the code architecture to isolate and optimizing GPU blocks that are executed in parallel. The focus will be on testing these blocks separately from the main framework and creating scripts which measure the performance of the blocks for different input parameters.

• Good command of Dutch or French. Can have a conversation in English. You will work in an international environment
• Knowledge of C++ (and Python), experience with object-oriented programming
• Knowledge of parallel programming or Nvidia CUDA is a plus
• Being able to work in a team, taking initiative in resolving issues and proposing solutions.
• Experience with larger software projects (multiple repositories, JIRA) is a plus
• Experience with cmake, gtest, Linux is a plus