Y02i10pdlzwdw8f0vt4f

HsienHao (Sam)  Hsu

I was an instructor in Dance Club, a theater assistant at Communication Department, and I was an international volunteer in Nepal. I believe that diverse background created diverse way of thinking. Which build up my core personality such as problem seeker and solver, enjoys challenge. Hope I can fully utilize data science skill and my personality to create value in the future

Education

Master of Science in Statistics and Data Science.2019 - 2020                           

University of Houston.GPA: 3.8/4

Related Courses:Data Visualization, Data Mining, Deep learning, Neural Network,Big Data Analytics

Tpknbzzw9iqzqlwclbsg

Exchange Program.2018 - 2019                                                                           

University of Connecticut

Related Courses: Project Management, Algorithm, Manufacturing Engineering, Business Analysis

Cupzslitt9mnvbwlh4nt

BS in Mathematical Sciences.Mathematical Finance Program.2014 - 2019

National Chengchi University

Related Courses: Data Structure, Financial, Investments, Financial Derivatives

Qg4v3wdkb4olzsjccexs

Work Experience

Big Data Development @ TransGlobe Life Insurance.2020/08 - now

• Analyze the main selling product changes after the Premium cliff on early July
Establish a machine learning model by python and obtain the analytical base table from daily report forms to predict high potential existing customers for cooperated agents

• Double the cooperated agents' sales turnover rate than the agents who didn't work with us

Bf8bdh2t4cw6dkw9phep

Project Management Team @ DHL Express.2017/06 - 2017/09

 Created 1700 interaction points by interactive activities of six sigma tools (DMAIC, 5S, etc.)

• Improve employees’ efficiency by established database for Import Customs Charge Department

Bf8bdh2t4cw6dkw9phep

Skills

Python (Numpy, Pandas, Scikit-Learn, TensorFlow, Keras) , R, C++, SQL, Spark, SAS, Tableau, Hadoop, AWS

Quantitative Projects 

Stock Price Predictor by ML 

• Built up a KRR model to predict stock price by other field relevant stock prices with 0.8% RMSE. Adjust the model by top 10 error cases.

Text Classification 

• Construct a 70% accurate rate classifier implement by Logistics Regression, Natural Language Processing to identify 2000 corpuses.

Digit Analysis Recognizer by Python 

• Apply KNN, SVM, PCA to identify tons of digits with 81.28% accurate rate

Marketing Advertisement Analysis 

• Formulate and visualize a marketing strategy by utilize EDA, K-Means and Hierarchical Clustering on customers' information to target market advertisement.

Deep Learning, Neural Network 

• Forecast stock price with 1.3% RMSE by using Auto Encoder technique and MLP predictor. Seek out the importance between each input and improve the algorithm.

Awards

1st PrizeTaiwan Society for Industrial and Applied Mathematics

Dual Information in Efficiency Analysis                                                   

• Improved about 20% efficiency of Chunghwa Telecom company by developing the dual problem of Data Envelopment Analysis; found the potential power for improving efficiency by dual, surplus and slack variables