Jia Hao Hu

Associate Researcher  •  Chunghwa Telecom Co., Ltd.

December 2016 - Present

• Created and maintained face feature extraction mobile SDKs for iOS and Android 
  • Refactored the code from MVP to the MVVM design pattern.
  • Utilized iOS/Android camera API and Tensorflow-lite framework to develop face recognition applications.
  • Used a convolutional neural network to train a facial landmark model to implement a liveness detection function.
    
  • Accomplishments 
    • Cooperated with the security team to develop the FIDO-certified biometric authentication SDK demonstrated at FinTech Taipei 2018. Link
    • Developed a recognition application for police patrol selected as the conceptual product at the Chunghwa 5G Demo Day.
• Developed and maintained a customer analysis platform for retailers to optimize their store layout.  
  
  • Responsible for front-end and back-end development based on the Spring framework.
    
  • Adopted the responsive design for better user experience across devices especially optimized for dashboard display.
  • Merged all image recognition services into the platform and provided related APIs.
  • Integrated Line/Telegram APIs to send alerts to store owners when specific users walk into the store.
  • Technique
    • OS: Linux (Ubuntu)
    • Environment: Tomcat
    • API: RESTful
    • Database: MariaDB, Redis
    • Monitoring:  Prometheus, Grafana
    • Test: JUnit 
    • Software: Docker
  • Accomplishment 
    • Performed a proof of concept with well-known retailers.
• Led the team to adopt Scrum and DevOps 
  • Worked with Scrum Master to identify the bottlenecks when running Scrum, define new execution processes,  and provide success cases for team members.
  • Applied Jenkins/Azure to execute the CI/CD processes including issue tracking, code control, build, test, release and deploy.
    
  • Leveraged Microsoft Azure platform to deploy microservices onto target machines on CHT Hicloud.