Po-chih Huang
Software Engineer at Microsoft
New Taipei, Taiwan
Summary
👤
Senior🎓
Top SchoolPo-chih Huang is a software engineer with 11 years of experience blending machine learning research, NLP, and full-stack web development, currently contributing to the Microsoft Advertising Platform. He holds a strong academic background from National Taiwan University (MS/BS, CS) and has hands-on research experience in deep learning and computer vision, including implementing FCN variants for semantic segmentation and building a content-based image retrieval system. Comfortable in both backend and ML roles, he has built data pipelines, training workflows, and production-ready components that bridge model research and application. Based in New Taipei, Taiwan, he brings a researcher's rigor to product engineering and a proven ability to move CV/ML prototypes into robust backend services.
11 years of coding experience
2 years of employment as a software developer
Master's degree, Computer Science, GPA 4.20/4.3, Master's degree, Computer Science, GPA 4.20/4.3 at National Taiwan University
Chinese, English
Github Skills (24)
pytorch10
preprocessing10
python10
evaluation10
preprocess10
semantic-segmentation10
histogram10
image-retrieval10
numpy10
data-preprocessing10
data-loading10
eval10
dataprep10
deep-learning10
model-testing10
Programming languages (3)
Github contributions (5)
pochih/FCN-pytorch
Oct 2017 - May 2018
🚘 Easiest Fully Convolutional Networks
Role in this project:
ML Engineer
ML EngineerContributions:60 commits, 53 pushes, 1 branch in 7 months
Contributions summary:Po-chih primarily focused on implementing and refining the Fully Convolutional Networks (FCN) architecture for semantic segmentation. Their contributions involved parsing and processing image data, defining the FCN model structure, including FCN32s, FCN16s, FCN8s and FCNs variants, integrating the VGGNet as the backbone, and setting up the training pipeline with a dataloader. The user also worked on splitting the training and validation data, and evaluating the model's performance using the Intersection over Union (IoU) metric.
pytorchsemantic-segmentationdeep-learningcomputer-visionconvolutional
pochih/CBIR
Nov 2017 - Sep 2020
🏞 A content-based image retrieval (CBIR) system
Role in this project:Backend Developer & Data Scientist
Backend Developer & Data ScientistContributions:75 commits, 2 PRs, 73 pushes in 2 years 10 months
Contributions summary:Po-chih primarily focused on implementing content-based image retrieval (CBIR) functionalities. Their contributions involved developing histogram-based feature extraction methods using Python and libraries like NumPy and SciPy. Key tasks included implementing and testing histogram calculations, defining distance metrics, and building a database interaction component. Furthermore, the user worked on feature extraction, experimenting with techniques related to computer vision and machine learning.
gaborcbircomputer-visioncontent-based-image-retrievaledges
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.