Mar 2022 - Present
• Researched and developed SOTA AI models for the industrial AOI using on Edge, Inspection, and cloud platforms.
• Built Edge AI software system for AI-integrated camera (ICAM product), including Edge AI model (Anomaly detection), front-end App, and back-end server for image capturing, AI training, and real-time inference. Developed 90% of the system's functions, and successfully led the development progress from inception to mass production within 4 months. Set up an exhibition at AI EXPO 2023 as the lead developer and reported directly to the VP.
• Developed features of an industrial AI AOI system (Inspection AI product), including data labeling, data augmentation, AI training, inference, and deployment. Identified and resolved a critical download and upload bottleneck between localhost and cloud storage, improving efficiency by up to 50%.
• Built and optimized responsive back-end systems using Sanic, zeroMQ, Docker, Boto3 (Amazon S3), etc. on multiple platforms such as Nvidia Jetson (Xavier AGX and NX), x86 ubuntu, and windows.