Intelligent Fabrication for Intelligent Vision: AI at the Core of CMOS Sensor Evolution

Abstract:

The convergence of advanced semiconductor manufacturing and artificial intelligence is redefining what’s possible in imaging and sensing technologies. CMOS image sensors (CIS) stand at the forefront of this transformation, driven by relentless demand for higher resolution, lower noise, and more intelligent pixel-level processing. This talk explores the critical role of semiconductor process innovation in enabling next-generation CIS, from 3D integration to innovative architectures.

We then delve into the transformative impact of AI on semiconductor manufacturing itself—highlighting how machine learning is reshaping process optimization, defect detection, yield prediction, and design-technology co-optimization. We further illustrate how AI-enhanced manufacturing pipelines not only improve efficiency and quality but also accelerate time-to-market for increasingly complex systems.

Finally, we outline key challenges and future directions at the intersection of AI and semiconductor manufacturing, offering a vision for how AI-driven methods will underpin the next decade of innovation in sensing technologies.

BIO:

Dr. Lorenzo Servadei is the Head of AI for Chip Design and Manufacturing at Sony AI. Currently he is lead of multiple teams of researchers, whose aim is the use of Machine Learning in the chip design and manufacturing process. Previously, he served as a Head of Machine Learning for Sensors at Infineon Technologies AG. In the same company, he earned his Ph.D. in computer science through collaboration with the Johannes Kepler University Linz. The topic of his Ph.D. thesis has been the development of AI algorithms for hardware-software co-design. His research outcomes have been published in several top conferences and journals such as DAC, DATE, TCAD, IROS, ICASSP. Parallel to his main occupation, he is a senior lecturer and group leader in Machine Learning for Design Automation at the Technical University of Munich, as a part-time “habilitand” in machine learning for hardware design at the chair for Design Automation