Graduate student in data science

EDUCATION

SKILLS LIST

PROJECT INTRODUCTION

LEADERSHIP EXPERIENCE

Xuewei Huang (Vivian)

213-***-**** ********@***.*** Los Angeles

www.linkedin.com/in/xuewei-huang-b7b5aa13a

University of Southern California 2017.08 - 2019.05 MSc Data Science Electrical and Electronic Engineering Los Angeles, USA GPA: 3.5 / 4.0

University of Liverpool 2015.09 - 2017.06

BEng Electrical and Electronic Engineering gineering Liverpool, UK GPA: 3.76 / 4.0

Xi'an Jiaotong-Liverpool University 2013.09 - 2015.06 BEng Electrical and Electronic En gineering Suzhou, China GPA: 3.76 / 4.0

Programming skills: Java, Python, C, C++, JavaScript, HTML, CSS, Bash, Git; Database: SQLite, MySQL, MongoDB; Computer Science Course: Data structure and Algorithm, Pattern Recognition, Natural Language Processing, Image Processing, Face Recognition, Computer Networks, Machine Learning, Deep Learning General skills: effective communication, problem solving and troubleshooting, critical thinking and writing Face recognition system design based on Keras and OpenCV 2019.01 - 2019.03 Face image capture and frame preprocessing in a video stream based on OpenCV with vector normalization and scaling Built a CNN model for training and used SGD algorithm to converge to give the initial prediction Trained KNN with k-fold cross validation to adjust CNN by changing softmax with 128-d representative features selected from Facenet

Obtained face images in the case of poor indoor lighting conditions at night and training accuracy rate reached 95.71% while test accuracy rate reached 99.63%

Android development design game similar to Pokemon Go 2018.12 - 2019.01 Designed a game similar to Pokemon Go in the environment of Android Studio Obtained users' location by listener and moving directions with Google Map API for real-time location system change and update Implemented SQLite database to realize the location storage and change, including energy storage and background information adding, removing, changing and retriving

Designed choice-able routines for user in case of real time traffic condition such as congestion optimization and weather prompts Twitter tweets party sentiment analysis based on Python 2018.10 - 2018.11 Crawled tweets and performed tag removing, link and image filtering, more deep text cleanup including emotion character substitution, case conversion, word vector embedding Applied TF-IDF to transform input into featured word frequency and inverse document frequency matrix Invoked different models training and tuning. Classic machine learning methods include naive Bayes, SVM, k-nearest neighbors, random forests, and deep learning methods like multi-layer perceptual NN and Keras-based CNN construction The forecast results are all above 65%, and CNN's results are best to reach 75% Java Web Development-Online Ticketing Recommendation System Design 2018.07 - 2018.08 Designed an event search ticketing site based on user preferences and used AJAX technology to implement interaction Created a Java servlet and use the TicketMaster API to handle HTTP requests and responses Built relational and non-relational databases for back-end support for real-time event recommendations (MySQL, MongoDB) Deployed the server to Amazon EC2 to handle 150 queries per second for Apache JMeter testing Kaggle mining and forecasting of a bank telemarketing dataset 2018.04 - 2018.05 Dataset preprocessing: missing value complemention, normalization, useless data reduction, and feature dimension reduction and adjustment (Python numpy, sklearn, pandas)

Randomize data sets and selected multiple model training, including perceptron, SVM, random forest and naive Bayes, and KNN, and use cross-validation to adjust parameters

The test results showed that 83% of customers will choose to continue to purchase wealth management products after once telemarketing

Calculated F1 scores and using the evaluation methods of fusion matrix, ROC, AUC, etc. to judge the training effect and prediction results of each model

Youth Volunteer Association 2014.05 - 2016.05

staff HR Suzhou

Coordinated with other departments to conduct Global Earth 1 hour activity as part of plotter and publisher Established daily human resource management and meeting arrangement as well as performance evaluation and collection

Contact this candidate