PhD - Leuven | More than two weeks ago
The progress of artificial intelligence requires ever increasing compute system (“exascale computing”), as demonstrated by recent industrial announcements (Dojo from Tesla and Ponte-Vecchio from Intel). As massively parallel computing units are integrated to meet these demands, these large systems become interconnect-limited. Photonic interposers have emerged as one of the most likely candidate technologies to solve the bandwidth bottleneck of these systems. Several high-tech start-up/scale-up companies (Lightmatter, Avicena) have recently announced that they are developing such photonic interposers, and diverse options are being considered. A systematic study combining photonics and system expertise is thus required to evaluate, compare and develop new photonic interposer concepts. In this context, the PhD candidate would investigate the following aspects:
Required background: Photonic, chip-to-chip communication
Type of work: Modeling (30%), Simulation (30%), communication with photonic/system experts (30%)
Supervisor: Wim Bogaerts
Co-supervisor: Nicolas Pantano
Daily advisor: Gaspard Hiblot
The reference code for this position is 2023-006. Mention this reference code on your application form.