Apr 2021 - Present
Develop, implement, and maintain platforms for inference engines. Develop and implement applications for automated optical inspection.
Major Accomplishments
● Porting OpenVINO inference engines to specific Intel-based devices on Windows and Ubuntu. Try to develop a system which is cross-platform and adapted to different inference engines
● Study GStreamer-related programming and combine it with OpenCV and OpenVINO for video streaming on Windows and Ubuntu
● Develop Automated Optical Inspection (AOI) applications for detecting screen defects using Python and Pyside2
● Develop a simple Optical Character Recognition (OCR) application using tesseract-OCR and QT
● Study various makefile systems, such as CMake, waf, and mason, to accomplish system integration on Windows and Ubuntu
● Survey LLMs(ChatGLM and LLaMa 2), fine-tunning, prompt engineering and langchain framework for building local ChatBot
● Use Ultralytics YoloV8/YoloV8-POS and Public dataset to train various models, including Face and Hand Detection with Landmark, License Plate Recognition and other object detection applications
● Survey Face Identify model GhostFaceNets, transfer and deploy to Intel OpenVINO Platform
● Survey Oneshot Object Detection Models
● Develop Speech Recognition Application using OpenVINO and OpenAI Whisper