Zxcfjm0pllq5klucgpcd

Laurence Lin

Enthusiastic data scientist, willing to bring up AI solutions that would solve human life problems. Highly cooperative teamwork and curious about data insight information. Familiar with Python and SQL language, experienced of image processing and time series forecasting. Understand basic web development and RESTful API with Flask or FastAPI framework.


Taichung city,Taiwan

Email: [email protected]

Cell Phone: 0916-067399

Github: https://github.com/laurence-lin

LinkedIn:https://www.linkedin.com/in/laurence-lin-836412133/

Medium: https://lawrence123.medium.com/

Working Experience


Machine learning engineer (2020/09/01 ~ current) 

Hamastar Technology Corporation

Build data science application platform for data science developers and end users 

Develop data science solutions for web layout defect detection, save 70% of time for developers to examine the product


Skills


Programming: Python, R, SQL 
Machine Learning: Regression, Classification, Time Series Forecasting 
Framework: Scikit-learn, Keras, PyTorch 
Web Development: html, css, RESTful API 
Web Scraping: Selenium 
Data Visualization: Tableau


Projects


Retail Store Location Selection by Popularity Ranking: Collect geographic features from Foursquare API and Google Place API, use EDA for data analysis and preprocessing, build predictive model to rank the store popularity, and found the main characteristics that effects most of the retail store's popularity. Achieve 0.83 in [email protected] metric comparable with previous research.

Methods: EDA, Google Place API, Foursquare API, SVR, Linear Regression, Neural Network

link: https://github.com/laurence-lin/Retail-Store-Location-Ranking


Credit Card Fraud Detection:  Predict the probability for credit card data fraud, apply feature engineering for data cleaning and transformation, use cross validation and ROC_AUC as evaluation for model performance.

Methods: EDA, LightGBM, Cross Validation, Logistic Regression

link: https://github.com/laurence-lin/Kaggle_competition/blob/master/FraudDetect/FraudDetection.ipynb

Projects


Retail Store Location Selection by Popularity Ranking: Collect geographic features from Foursquare API and Google Place API, use EDA for data analysis and preprocessing, build predictive model to rank the store popularity, and found the main characteristics that effects most of the retail store's popularity. Achieve 0.83 in [email protected] metric comparable with previous research.

Methods: EDA, Google Place API, Foursquare API, SVR, Linear Regression, Neural Network

link: https://github.com/laurence-lin/Retail-Store-Location-Ranking


Credit Card Fraud Detection:  Predict the probability for credit card data fraud, apply feature engineering for data cleaning and transformation, use cross validation and ROC_AUC as evaluation for model performance.

Methods: EDA, LightGBM, Cross Validation, Logistic Regression

link: https://github.com/laurence-lin/Kaggle_competition/blob/master/FraudDetect/FraudDetection.ipynb

Education


IBM Data Science Professional(Certified) 

Coursera, June 2020 ~ August 2020

Google Data Analytics(Certified)

Coursera, April 2021 ~ June 2021

National Cheng Kung University / Engineering Science / Bachelor's degree

國立成功大學工程科學系

National Cheng Kung University / Engineering Science / Master degree

國立成功大學工程科學所

Master thesis: PM 2.5 forecasting using LSTM model


Powered by CakeResumePowered by CakeResume