NTHU Human-Centered Machine Intelligence (HMI) Lab Advised by: Ph.D. Po-Chih Kuo (郭柏志) Assisted in building computer vision ML models for applications in the medical field. Using 𝗖𝗡𝗡, and 𝗬𝗢𝗟𝗢 models with 𝗳𝗶𝗻𝗲-𝘁𝘂𝗻𝗶𝗻𝗴 for optimal performance towards the lab's use-case. 𝗥𝗼𝗹𝗲𝘀: Machine Learning Engineer

CS235100 (Chen, Yi-Shin / 陳宜欣) EECS121000 (Chou, Pai-Hsiang / 周百祥)

Took ownership in pioneering AI initiatives at OMISE, working with several 𝗔𝗪𝗦 𝘀𝗲𝗿𝘃𝗶𝗰𝗲𝘀 (Bedrock, S3, IAM). Utilized 𝗥𝗔𝗚, and 𝗟𝗮𝗻𝗴𝗖𝗵𝗮𝗶𝗻 framework. The solution is a proof-of-concept customer service AI agent that help client onboard and integrate to company's products seamlessly. The agent also offers interactive chat via custom 𝗠𝗖𝗣 𝗽𝗹𝘂𝗴-𝗶𝗻𝘀. • 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗥&𝗗 - researched 2 potential technologies for agentic workflows (n8n, Bedrock). • 𝗗𝗲𝘀𝗶𝗴𝗻𝗲𝗱 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗪𝗼𝗿𝗸𝗳𝗹𝗼𝘄 - following enterprise's design patterns for compatibility to existing ecosystem. • 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗱 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 - bringing 3 new tech stacks and solutions to the company's AI initiatives. • 𝗠𝗖𝗣 𝗜𝗻𝘁𝗲𝗴𝗿𝗮𝘁𝗶𝗼𝗻 - using a custom MCP tool to enhance AI Agent's capabilities. • 𝗦𝗵𝗶𝗽𝗽𝗲𝗱 𝘁𝗼 𝗨𝗔𝗧 - through 𝗗𝗼𝗰𝗸𝗲𝗿 deployments, 𝗞𝘂𝗯𝗲𝗿𝗻𝗲𝘁𝗲𝘀, and 𝗧𝗲𝗿𝗿𝗮𝗳𝗼𝗿𝗺. Released for internal testing. 𝗥𝗼𝗹𝗲𝘀: AI Product Development, DevOps Engineering, Full-Stack Development

Collaborated with a team of developers to build, test, and deliver 𝗖𝗫𝗕𝗢𝗫, a middleware product designed to enhance customer servicing experience on multiple e-commerce platforms across APAC. Our product allow customer service staff to efficiently communicate with customers across various online shopping platforms in one place! • 𝗗𝗲𝘃𝗲𝗹𝗼𝗽𝗲𝗱 𝗮𝗻𝗱 𝗘𝗻𝗵𝗮𝗻𝗰𝗲𝗱 𝟯 𝗡𝗲𝘄 𝗙𝗲𝗮𝘁𝘂𝗿𝗲𝘀: For product website and application, leading to a more intuitive user experience. • 𝗘𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝗲𝗱 𝗡𝗲𝘄 𝗔𝗣𝗜 𝗠𝗲𝘁𝗵𝗼𝗱: To Sync database with application's AI-based feature. • 𝗦𝗼𝗳𝘁𝘄𝗮𝗿𝗲 𝗧𝗲𝘀𝘁𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗮𝘁𝗶𝗼𝗻 𝗧𝗼𝗼𝗹𝘀: Created automated test units to identify issues, ensuring project stability. 𝗥𝗼𝗹𝗲𝘀: Front-End Development, API Development, Software Quality Assurance