Skills
• Languages: Python
• DataBases: MySQL, SQLite
• Infrastructure tools: Github
• Machine learning libraries: TensorFlow, Keras, and Scikit-learn
• Data visualization tools: Power BI, Seaborn and Matplotlib
• Deployment: Streamlit
Summary
I have been working in Motor Manufacturing Industry for 8 years. My first programming was going to my Bachelor's degree, C++ was the first program I learned. Then I started to learn Python in 2018 at TEDU and my first project was the Stock Trend Prediction by CNN. I kept using Python to implement web crawling, OOP, and Pandas in my job, intend to let my work become more automated. I used those techniques to automate the data-gathering problem, which shorten the existing progress duration. I'm very passionate about Data Scientist and Machine Learning.
Work Experience
Product Manager • 東元電機股份有限公司 (TECO Electric & Machinery Co. Ltd.)
January 2023 - October 2023
1. Product Analytics
2. Market Trend Analytics
3. New Product Development Project Management
4. Industrial Induction Motor Instructor
Team Lead Supervisor • 東元電機股份有限公司 (TECO Electric & Machinery Co. Ltd.)
May 2021 - December 2022
1. Industrial Induction Motor Designer
2. R&D
3. Instructor of Oversea's Quotation and Engineering Department
4. Project Leader
-- Leading a team with 5 people
Senior Researcher • 東元電機股份有限公司 (TECO Electric & Machinery Co. Ltd.)
January 2020 - April 2021
1. Industrial Induction Motor Designer
2. R&D
3. Instructor of Oversea's Quotation and Engineering Department
4. Project Leader
Assistant Researcher • 東元電機股份有限公司 (TECO Electric & Machinery Co. Ltd.)
February 2015 - December 2019
1. Industrial Induction Motor Designer
2. R&D
3. Quotation and Engineering Consultant
4. Instructor of Oversea's Quotation and Engineering Department
Side Project
Genetic Algorithm Electrical Motor Design Program
This project is going to establish a program that can maximize the power density of electrical motor. This program can automatically design an electrical motor that has a lighter weight but the highest efficiency.
Initialization: The motor design parameters become the genetic parameters, and motor characteristics will be set as the constraint parameters.
Selection: Since the motor characteristic is a discrete type of data, therefore, fuzzy logic is carried out to define how good this chromosome is, which is the calculation of fitness. I use Deterministic Sampling for the first round of selection and then the Enlarged Sampling Space method for the rest generation. The Elitism Selection has also been used in the Selection Stage.
Genetic Operator: Double Crossover is the method that is used in the crossover operator.
- Skills： Matlab · Genetic Algorithms
Pricing Web Crawling Program
This program solves the pricing update problem such as raw materials, accessories, and other parts. By combining the EFC Calculation Program, the complicated workload for every month's update can be more convenient.
- Skills：Pandas (Software) · Web Crawling · OOP · Python
Stock Prediction by CNN
This project is about doing the stock prediction by CNN and using the triple barrier method to do the labeling.
Frontend: Streamlit
Crawling: Gather the data from the website by downloading the data or xpath to work on the data.
Data Preprocessing: 15 kinds of technical indicators are used to transform to a 15x15 pic, then drop those NA columns. The triple barrier method was used to do the labeling by inputting the upper bound of the stop profit and the lower bound of the stop loss, by the way, the holding period can be the input parameter also. Normalizing was used before transforming the data into a pic. by MinMaxScaler.
Training: I establish two different models which are the normal CNN model with RELU activation function and categorical cross-entropy with ADAM as optimizer to calculate the loss. The other one is using LeakyRELU as the activation function and RANGER optimizer in categorical cross-entropy loss calculation.
Finally, I will use the financial evaluation to calculate the profit I earn by using these two models.
- Skills：Pandas · Web Crawling · OOP · Github · Convolutional Neural Networks (CNN) · Python
Education
國立成功大學 National Cheng Kung University
Mechanical Engineering
Master Degree, 2012 - 2014
國立成功大學 National Cheng Kung University
Mechanical Engineering
Bachelor Degree, 2008 - 2012
Licenses and Certifications
TOEIC
ETC - Score: 855
Python Data Analytics Course
上台科技股份有限公司（達內教育TEDU）
210468000009668833
Issued April 2021 ·
Data Science: Modern Deep Learning in Python
Udemy
UC-8e281a65-b219-4c80-84fd-080d64600072
Issued May 2023 ·
The Python Mega Course: Learn Python in 50 Days with 20 Apps
Udemy
UC-9bad5f82-d2b4-45c6-9b18-9d65e64f458f
Issued March 2023 ·
Data Science: Deep Learning and Neural Networks in Python
Udemy
UC-76281cb5-4575-43a2-abaf-32271e5d4a0f
Issued March 2023 ·
數據思維課程
ALPHA CAMP
Issued February 2024 ·
Beginner's Guide to Power BI
ALPHA CAMP
Issued February 2024 ·
Top Skills
Python Power BI Data Analytics Machine Learning Induction Motor Theory Web Crawling Marketing Strategy
Languages
English — Fluent Chinese — Native or Bilingual Malay — Fluent
