* Designed application software products such as - Offline data analysis tool- Online report generation tool- Recipe editor for both single-beam and multi-beam machines* Unit Testing* Troubleshooting

* Developed the line width and pitch measurement Winform application using C#. Compared to the laser-based solution, it only takes 3% of measurement time per piece. * Established a concurrent measurement algorithm. After the performance enhancement, it takes only 17% of the time to measure the mask compared to the previous way.* Trained convolutional neural network (CNN) models to execute binary classification using TensorFlow. * Tuned the CNN model performance with less than 20 ppm skip rate so that it can fulfill the specification of our client. * Added feature to the training model procedure to raise the accuracy using Python. * Maintained the defect classification and defect detection software which is built by C# and WPF.

* Contributor of open source project SOFA (Github: https://github.com/cyliustack/sofa)* Designed an easy-to-use module to cluster countless data and target the potential performance bottleneck* Research Assistant of Performance, Applications, and Security Lab (PASLab), Computer Science and Information Engineering (CSIE), National Taiwan University (NTU)

* Teaching Assistant of Parallel Computing, a graduate school course of CSIE. * Responsible for 50% of the homework assignment, covering up the knowledge of parallel computing, computer vision, and TensorFlow.* Coworking with team members to host the project assignment.

* Host final project of material mechanics

* Teaching vocational high school math and material mechanics