• Built a rtsp server for the medical endoscope video source in NVidia TX1 using gstreamer library and C++ which enables the customer to access the 1280*1080 and 10frame/s video from remote device• Built the Tensorflow platform in NVidia TX1 develop kit and developed the action recognition model for a kindergarten in Fremont which can detect running, walking, eating, etc for the kids. Keras Python library and jupyter are used in development.• Developed and maintained the Qt application software of the endoscope system using C++ and Qt API, which contains the GUI and the backend function such as video save function, video zoom function, etc• Built the real-time video object detection in the Nvidia tx1 system using YOLO algorithm and increase the detection accuracy to 93% for 10 frame/s video sources which applied to security alarm system.

•Developed a software based on C++ in Qt and arduino to do the sensors calibration.•Built the cloud database for the sensors’ calibration data.•Developed an IOS app for the customers to search their information and data.•Maintaining the company's website. •Developing a software based on C++ in Qt and arduino to track realtime pressure distribution in a chair.