The Fraunhofer-Gesellschaft (www.fraunhofer.com) currently operates 76 institutes and research units throughout Germany and is a leading applied research organization. Around 32 000 employees work with an annual research budget of 3.4 billion euros.
Developing innovative technology solutions and bringing them to application - that is our goal at the Fraunhofer Institute for Photonic Microsystems IPMS. With our expertise in the development of photonic microsystems, related technologies including nanoelectronics and wireless communication solutions, we create - in flexible and interdisciplinary teams - technologies for innovative products in a wide range of markets such as automotive, industrial and aerospace.
At the Center for Nanoelectronic Technologies (CNT) of Fraunhofer IPMS, we focus on developing advanced AI systems through the integration of both Convolutional Neural Networks (CNNs) and Spiking Neural Networks (SNNs). This position will involve training and optimizing these neural networks using Python frameworks, including CUDA, Keras, PyTorch, and SNN-specific libraries like SNNtorch. The role will also explore hardware-aware design strategies such as quantization, error resilience, and co-training techniques for optimized performance on neuromorphic hardware.

• Develop and train CNN and SNN models using frameworks like Keras, PyTorch, and SNNtorch
• Implement GPU acceleration using CUDA for efficient training of neural networks
• Explore hardware-aware design concepts such as quantization, error injection, and mixed-precision training
• Design co-training methods to improve model resilience to hardware imperfections and failures
• Analyze model performance, including the impact of hardware-aware strategies on efficiency and accuracy
• Collaborate with other researchers to adopt neural network models for hardware-specific optimizations
• Document the research process, experimental results, and best practices for future projects