Experience

McWilliams School of Biomedical Informatics at UTHealth Houston

• United States
• Higher Education
• 1 - 100 Employee

• Graduate Research Assistant

  • 2022 - Present

  ◦ Developed deep learning models to forecast protein-ligand binding affinities with Pytorch, achieving 18% higher accuracy for 16 different ligands when compared to traditional methods. ◦ Gathered, standardized, and analyzed more than 500,000 protein-ligand 3D structures and their binding affinity data from 6 disparate sources. ◦ Transformed raw data into compelling visual representations resulting in a 30% increase in audience engagement and understanding of complex data concepts. ◦ Developed deep learning models to forecast protein-ligand binding affinities with Pytorch, achieving 18% higher accuracy for 16 different ligands when compared to traditional methods. ◦ Gathered, standardized, and analyzed more than 500,000 protein-ligand 3D structures and their binding affinity data from 6 disparate sources. ◦ Transformed raw data into compelling visual representations resulting in a 30% increase in audience engagement and understanding of complex data concepts.



INNOLUX群創光電

• Taiwan
• Appliances, Electrical, and Electronics Manufacturing
• 700 & Above Employee

• Research Engineer

  • 2020 - 2021

  ◦ Implemented optimization algorithms that facilitated material selection and processing parameters tuning, increasing product yield rate by 27%. ◦ Applied finite element analysis software to simulate temperature and mechanical stress distribution for IC package processes, saving product development schedule by 20%. ◦ Established a MySQL database including 25 distinct materials and their mechanical/thermal properties, bringing 40% time reduction in the material selection process. ◦ Optimized material property measurement processes and further reduced measurement time by 5+ hours per week. Show less



National Cheng Kung University

• Taiwan
• Higher Education
• 700 & Above Employee

• Graduate Research Assistant

  • Sep 2018 - Aug 2020

  ◦ Designed machine learning models which predict molecular chemical properties, achieving a 28% decrease in RMSE than previously published methods. ◦ Implemented feature selection techniques to identify the relationship between chemical structures and chemical properties, promoting electrolyte molecular design. ◦ Developed and streamlined a high-throughput pipeline to validate molecules in Python and Shell Scripting, resulting in an 80% reduction in validation time. ◦ Collected, integrated, and processed over 6 million molecular structural and their related physical properties data. Show less



National Sun Yat-Sen University

• Taiwan
• Higher Education
• 400 - 500 Employee

• Undergraduate Research Assistant

  • Jul 2016 - Aug 2018

  ◦ Conducted first-principles calculation to investigate of the hydrogen evolution reaction on different surfaces of pyrite material. ◦ Investigated the electronic interaction between transitional metal atoms and hydrogen atoms to illustrate the discrepancy of HER efficiency on different surfaces and different materials. ◦ Automated time-consuming calculation jobs and material model-building tasks with Python code. ◦ Conducted first-principles calculation to investigate of the hydrogen evolution reaction on different surfaces of pyrite material. ◦ Investigated the electronic interaction between transitional metal atoms and hydrogen atoms to illustrate the discrepancy of HER efficiency on different surfaces and different materials. ◦ Automated time-consuming calculation jobs and material model-building tasks with Python code.