2020 - 2022
2016 - 2020
July 2021 - September 2021
Taipei, Taiwan
Develop two programs from scratch in 10 weeks including:
1. Big Data collection tool (C++)
◾ collect 80000+ device information in 5 seconds
2. Production information prediction system (Python)
◾ leverage state-of-the-art Machine Learning algorithms
◾ Database connection
A fast and lightweight autoML library developed by Microsoft Research
Predict Future Sales
Rank: 20/12544
A regression task in structured data
Medical data De-identification
Rank: 36/531
A Name Entity Recognition task in medical conversation data
Date: 2018/05/29
Problems solved: 4/7
Rank: 110/2181
Python
C/C++
Java
C#
R
Neural Network
Transfer Learning
Ensemble Learning
Self-Training
Image Process
Image Classification
Object Detection
Object Tracking
Hadoop Docker OpenCV tensorflow Scikit-Learn Windows API Ubuntu git
◾ Object Detection: fine tune pretrained models from tensorflow zoo include SSD ResNet and CenterNet HourGlass architectures.
◾ Object Tracking: Implement a custom object tracking API including features like deepsort and non-maximum suppression.
◾ Real-time data display
◾ Real-time outlier detection
◾ User Database
◾ Implement image process algorithms from scratch
◾ leverage OpenCV API in image/video process
◾ Train high accuracy models over 7+ image datasets
◾ Strong knowledge in Transfer learning and Neural Network training details
◾ Multi-user chat room
◾ Unzip file server (Huffman Coding)
◾ TCP congestion control simulation
2020 - 2022
2016 - 2020
July 2021 - September 2021
Taipei, Taiwan
Develop two programs from scratch in 10 weeks including:
1. Big Data collection tool (C++)
◾ collect 80000+ device information in 5 seconds
2. Production information prediction system (Python)
◾ leverage state-of-the-art Machine Learning algorithms
◾ Database connection
A fast and lightweight autoML library developed by Microsoft Research
Predict Future Sales
Rank: 20/12544
A regression task in structured data
Medical data De-identification
Rank: 36/531
A Name Entity Recognition task in medical conversation data
Date: 2018/05/29
Problems solved: 4/7
Rank: 110/2181
Python
C/C++
Java
C#
R
Neural Network
Transfer Learning
Ensemble Learning
Self-Training
Image Process
Image Classification
Object Detection
Object Tracking
Hadoop Docker OpenCV tensorflow Scikit-Learn Windows API Ubuntu git
◾ Object Detection: fine tune pretrained models from tensorflow zoo include SSD ResNet and CenterNet HourGlass architectures.
◾ Object Tracking: Implement a custom object tracking API including features like deepsort and non-maximum suppression.
◾ Real-time data display
◾ Real-time outlier detection
◾ User Database
◾ Implement image process algorithms from scratch
◾ leverage OpenCV API in image/video process
◾ Train high accuracy models over 7+ image datasets
◾ Strong knowledge in Transfer learning and Neural Network training details
◾ Multi-user chat room
◾ Unzip file server (Huffman Coding)
◾ TCP congestion control simulation