Sort by
Refine Your Search
-
Listed
-
Country
-
Program
-
Field
-
Ryzen CPU, GPU, and NPU, in terms of inference speed, energy consumption, accuracy, and performance per watt. Different quantization levels (e.g., int8, fp16) will also be explored. Develop intelligent
-
are a core member of IMEC, the world-leading research and innovation center in nanoelectronics and digital technologies. Our team is currently a fertile mix of people of different nationalities. The main
-
optimizing compilers, the classical and quantum fragments are separated in efficient implementations adapted to the changing QPUs and GPUs architectures. The candidate will work at the intersection
-
of spikes by a model Develop proxy apps representing the different processing stages of spiking network simulation code (targeting CPU and accelerators such as GPU or IPU) Systematic benchmarking of proxy
-
of advanced language models and derived use cases by focusing on one or more of the following topics in their PhD project: Training and inference of ML models on GPU clusters. Method development for scalable
-
-design with accelerators (FPGAs, GPUs, near-memory systems) to achieve real-time, energy-efficient AI for high-tech industry applications. Work with leading companies like ASMPT and shape the future of AI
-
publications at top-tier venues such as CVPR, ICCV, ECCV, NeurIPS or ICRA. You will have access to extensive compute resources at TU Delft, ranging from local GPU servers to large-scale HPC infrastructure
-
algorithms Extend the superstructure to tackle AC-PF problems of different complexities and assess its convergence in inference Investigate scaling and performance bottlenecks Explore hybrid ML-classical
-
to the development of advanced language models and derived use cases by focusing on one or more of the following topics in their PhD project: Training and inference of ML models on GPU clusters. Method development
-
. Your tasks in detail: Become familiar with our previously developed neural network superstructure for learning iterative algorithms Extend the superstructure to tackle AC-PF problems of different