Yuanfu Yang

Senior Software Engineer

TSMC•  September 2014 - Present

 Create data-driven solutions to semiconductor operations problems (e.g. defect detection, fail rate analysis) across a variety of department using machine learning and statistics.  Used EDA, statistical tools, machine learning modeling algorithms and model performance tools for creating a backend model to estimate control wafer fail rate.  Researched and implemented computer vision based defect detection and provided continual support for product.  Developed and deployed deep learning models for object detection.  Built statistical models to recognize risky situations using a combination of inline and offline monitor in foundry.  Improved existing models by resolving practical and algorithmic issues.  Develop scalable production Python code for client deployment.  Support for Data Science project.

Researcher

National Tsing Hua University•  September 2019 - Present

Researcher at artificial intelligence program.  Projects: * Cross-lingual AI and DH Study on Semantics, Discourse, and Pragmatics - AI Poet. * Covid-19 Detection by X-ray Image. * The Rapid Spread of COVID-19 Prediction. * Wafer Map & Pattern Defect Detection.  Publication of Works: “A Novel Deep Learning Architecture for Global Defect Classification: Self-Proliferating Neural Network (SPNet)”, 2021 Semi Advanced Semiconductor Manufacturing Conference (ASMC 2021). “Double Feature extraction method for wafer map classification based on convolution neural network”, 2020 Semi Advanced Semiconductor Manufacturing Conference (ASMC 2020). “A Deep Learning Model for Identification of Defect Patterns in Semiconductor Wafer Map”, 2019 Semi Advanced Semiconductor Manufacturing Conference (ASMC 2010).

Software Engineer

TSMC•  July 2011 - August 2014

 Design a dashboard, a frontend UI development, as a visualization tool for the user.  Responsible for the planning and re-designing the internal work flow of the semiconductor manufacturing line. Achieved a productivity improvement of 40% with a auto workflow built in Web UI.  Wrote simulations to study the sensitivity of equipment productivity.  Use a variety of Big Data tools to analysis semiconductor process history and solve the system problem in foundry.