via XGB model. ● benefit: Expected benefit is 280,000 NTD/year. 3. PHM - Find faults of water supply pump quickly via Logistic regression. ● benefit: 99% prediction accuracy and maintain the stability of production. Responsible for: ● Improve factory production with machine and deep learning ● Help the development of smart manufacturing projects and the parallel development of the factory (Equipment warning, energy saving and fee reduction, etc.) Data Analyst • Digiwin Software Co., Ltd. 九月九月 2020 Project achievements: 1. Predictive Maintenance and Quality ● content: Calculated

incoming mails every week using Text vectorization (NLP) and SGD classifier. - Identifying 78% of the booking and voucher frauds using a random forest anomaly detection model. - Created a RASA chatbot to help customers find answers to booking related questions and requests. Data Scientist • Fraunhofer, Paderborn FebruaryJulyDeep Q-learning to train an autonomous car to navigate in an Industry 4.0 environment. - Statistical and deep learning models for low latency (~3.8s) anomaly detection on acquired sensor data. - Developed, tested, deployed, and managed continuous delivery and automation pipeline for this solution. - Interactive visual