Deep-Learning Assisted Laser Chip Coupling for Hybrid Integrated Photonics

Hybrid integration is a promising technology for Photonic Integrated Circuits (PIC). This would allow the combination of different technological platforms in a single package for emerging applications in AI data centres, lab-on-a-chip, microwave photonics and quantum technologies [1], [2]. In hybrid-integrated photonics, micro-lenses are used to couple active laser chips to passive integrated chips and optical fibers either for testing or final packaging. In this work, we present a novel deep-learning approach for automated coupling of Quantum-Dot (QD) multi-strip laser chips to micro-lenses.