Experience

Fitipower Integrated Technology Inc.

• Taiwan
• Semiconductors
• 1 - 100 Employee

• AI algorithm manager

  • May 2023 - Present

  1. Lead the team to efficiently produce deliverable AI models. ==>My method trains the model 10x faster and needs 3x less data than the standard approach. 2. Guide a team member to develop an AI model, under 500KB in size, to a level comparable with industry leaders, for client demos within three months. 3. Skilled at efficiently leading team members to successfully develop high-quality AI models. 4. Optimize the AI model to fit the hardware architecture. 5. TinyML AI algorithms development. Show less



Altek Corporation

• Taiwan
• Computers and Electronics Manufacturing
• 1 - 100 Employee

• CV/AI演算法技術經理

  • Apr 2021 - May 2023

  1. Have over 15 years of AI and CV experience. 2. Optimize AI models to run on edge devices. (Have practical edge AI productization experience) 3. Developed real-time hand, body, and facial tracking models with 60 fps or higher on a low-power chip. 4. Trained AI model for high accuracy with a very small dataset (<3K). 5. Use a single small model to achieve multiple functions for saving hardware computing resource. For example, image classification, object detection, and image segmentation all are performed simultaneously by a model that is smaller than 0.5 MB, and achieve real-time performance on edge chips. 6. The developed model allows users to define and add new recognition/segmentation categories on their own without needing to retrain the model. 7. The developed model can understand the real-world application environment and improve its accuracy on its own even after leaving the factory. 8. Developed an SOP for quickly generating feasibility assessment models, and typically within two weeks, we can determine whether this method can run on the target device as expected. This significantly boosts development efficiency and cost-effectiveness. 9. Developing AI models to remove redundant information from continuous video frames, resulting in a significant reduction in power consumption and increased inference speed of the AI models. 10. Plan/design competitive AI new product features. 11. Managed the progress of AI projects. 12. Propose a method that can effectively avoid data imbalance. 13. Lead the AI + ISP team. a. Improve image quality based on AI technology. b. Real time video semantic and instance segmentation development. Segment image into meaningful regions and improve image quality by processing each region according to its category and related information. Show less



Foxit

• United States
• Software Development
• 200 - 300 Employee

• Computer Vision Scientist

  • Sep 2020 - Mar 2021

  Develop intelligent document technology by advanced technologies, such as computer vision and deep learning. Develop intelligent document technology by advanced technologies, such as computer vision and deep learning.



HTC

• Taiwan
• Consumer Electronics
• 700 & Above Employee

• Senior Manager

  • Dec 2018 - Aug 2020

  Develop deep-learning based algorithm using a single regular RGB camera, such as object recognition, object detection, 3d hand/body pose tracking, object segmentation, depth estimation, scene recognition, gesture recognition and human action recognition.Product:[Vive Facial Tracker]Videos:1. https://www.youtube.com/watch?v=l_l4_aoiVAw2. https://www.youtube.com/watch?v=MPcXPJzwricPatent: Facial expression modeling method, apparatus and non-transitory computer readable medium of the same( https://patents.google.com/patent/US10885702B2/en?inventor=Cheng-Hsien+Lin&assignee=htc&oq=inventor:(Cheng-Hsien+Lin)+htc ) Show less

• Manager

  • Jun 2017 - Dec 2018

  Develop deep-learning based algorithm using a single regular RGB camera, such as object recognition, object detection, 3d hand/body pose tracking, object segmentation, depth estimation, scene recognition, gesture recognition and human action recognition.

• Assistant Manager

  • Jun 2015 - May 2017

  Achievements:1. Developing Al-related applications about image recognition, object detection, gender recognition, scenesunderstanding, and so on.2. Make-up effect development, including skin lighting, eye brightening, face beauty...3. Ufocus application development4. Image quality improvement: image sharpness, blur detection, lighting, noise removal, auto scene detection...

• Principal Engineer

  • Jun 2013 - May 2015

  Algorithm development for smart phone camera, image effects and quality.

• Senior Computer Vision Engineer

  • Nov 2011 - May 2013

  Algorithm development for smart phone camera, image effects and quality.



由田新技股份有限公司

• Appliances, Electrical, and Electronics Manufacturing
• 1 - 100 Employee

• Computer Vision Engineer

  • Sep 2010 - Oct 2011



huperLab

• Taiwan
• Software Development
• 1 - 100 Employee

• Computer Vision Engineer

  • Aug 2009 - Sep 2010



啟碁科技

• 台灣 Taiwan 新竹市

• 手機應用程式軟體工程師

  • Feb 2006 - Sep 2007

  手機應用程式軟體工程師 手機應用程式軟體工程師