Yuki Shizuya

• I have worked with manufacturing industry clients to improve the efficiency of quality assurance. To be concrete, my client examined product quality manually, and it took time. Our team created the deficient finding object detection model, and this model could reduce 60% of its process. I was in charge of preprocessing, data modeling, and the backend of the Web App.
• I have worked with logistics industry clients to predict truck drivers' behavior. Currently, they have a problem in that shippers have massive power than logistics companies, and shippers force them to work additional work without payment. Our teams used smart devices to obtain motion information to unveil that fact. Then we created the behavior classification AI model to determine whether the target driver records the correct duty status.
• I have experience in building an on-premise machine learning analysis server. I was in charge of the network, security, GPU setting, and docker environment. Moreover, I have built an MLflow server by Cloud Run to share experiment results quickly and easily.

・I have created an environmental-friendly food packaging film for three years. I replaced 25% of the significant clients' film using bio-polyethylene.

・I have analyzed the polymer structure when heating them using fluid mechanics.

・I have created a database to gather experimental data and a machine-learning model to predict material properties before experimenting.