Data Science Manager, LAM Research
ArulMurugan Ambikapathi is a Data Science Manager at Lam Research Singapore, responsible for advanced deep learning-based computer vision and Inverse Design applications in Semiconductor manufacturing. Prior to that, he was a Group Lead and Scientist in Deep Learning 2.0 / Machine Intellection, Institute of Infocomm Research, a research wing of Agency for Science, Technology, and Research, Singapore, from 2018 to 2022, and was focusing on semiconductor industrial research and smart AI applications. From 2014 to 2018, he was a Project and Team lead at Utechzone Co. Ltd., Taipei, Taiwan, where he developed one of the first AI based robust and precise Defect Inspection machines, for several manufacturing sectors.
He had received his Ph.D. degree from the Institute of Communications Engineering (ICE), National Tsing Hua University (NTHU), Taiwan, in 2011. During his PhD and Postdoctoral research (2011-2014), he has been focusing on advanced 3D (hyperspectral images) / 4D (DCE-MRI) Image analysis. He had also received his master (University Top 3 ranks - 2005) and bachelor (Gold Medalist - 2003) degrees in communication engineering with signal and image analysis as the major. His earlier research expertise includes convex optimization, biomedical and hyperspectral image analysis. His current research and application interests are in advanced computer vision, Bayesian optimization, and online / continual learning (theories and applications). He has 7 US / China / Taiwan Patents and more than 50 publications in top-tier Conferences and Journals.
Presentation Title: "AI Era" for End-to-End Process Development
AI has just begun to revolutionize the semiconductor industry ranging from material identification, process and yield optimization, to supply chain sustainability, and even human resource management. We are in an important pivot point leading into this transition. In this talk, I will be focusing on the spectacular role that AI can play in one of the core areas of semiconductor manufacturing, which is end-to-end process development. Specifically, I will discuss the most recent innovations in AI-based process recipe development and DL/ML-based image metrology to obtain robust and precise nm level / subpixel level measurements. These two components, recipe development and metrology, are highly interdependent and making them "smart and reliable" is vital to enable rapid speed to solution. Will conclude this talk by establishing the key components' connections and discussing future directions.