(Bryan) Zeyang Yu
Jersey City, NJ
ac0f1r@r.postjobfree.com Phone: 484-***-****
LinkedIn: https://www.linkedin.com/in/zeyang-yu-machine-learning/
SUMMARY
Data Scientist with project experience specializing in machine learning, predictive modeling, relational database query, natural
language processing, experimental design, data structure and algorithm using Python (and its libraries Scikit-learn, Pandas, NumPy,
TensorFlow), SQL and VBA. Proven skills in working with large data sets (Hadoop), A/B testing, building convolutional/recurrent neural
network. Proficient in machine learning techniques such as neural network and linear regression and their application in business.
EDUCATION BACKGROUND
Udacity, Deep Learning Nanodegree, Expected June 2017 Online Degree
Udacity, Machine Learning Nanodegree, Expected June 2017 Online Degree
Lehigh University, M.Eng. Industrial and System Engineering, May 2016 (GPA: 3.5/4.0) Bethlehem, PA
Jilin University, B.S. Automotive Engineering, May 2014 (GPA: 3.7/4.0) Changchun, China
SKILLS
Tools: Python (Scikit-learn, Pandas, NumPy, TensorFlow), SQL, VBA, SAS
Technical: Random Forest, Logistic Regression, SVM, PCA, Convolutional/Recurrent Neural Network, Hadoop/MapReduce
PROJECT EXPERIENCE
Apply PCA and K-means Clustering to make customer segmentation, Udacity May 2017
Applied Scikit-learn in Python to make customer segmentation using data on customers' annual spending amounts of diverse
product categories, to help the distributor on how to best structure the delivery service to meet the needs of the customer
Conducted feature scaling, outlier detection, then performed feature transformation (PCA) to reduce the dimension of the
data from six to two
Implemented K-means clustering method on two principle components and separated the customers into three segments
(with silhouette coefficient = 0.412)
Apply Supervised Algorithms to Find Donors for Charity, Udacity April 2017
Implemented different machine learning algorithms (Random Forest, Logistic Regression, Support Vector Machine) with scikit-
learn to predict whether an individual makes more than $50,000 a year
Conducted data preprocessing to make the data cleaned and formatted (one-hot encoding), then trained different models and
evaluated their performance
Selected Logistic Regression as the best model, then performed feature transformation(PCA) and grid search to achieve 85% of
accuracy and 75% of F-score
Bike Sharing Prediction with Neural Network, Udacity March 2017
Built a neural network from scratch to predict daily bike ridership
Implemented gradient descent and backpropagation with NumPy and Pandas to train the neural network
Adjusted the hyperparameters (iterations, learning rate, hidden nodes) and analyzed bias/variance tradeoff to find the optimal
point of minimizing training loss and validation loss while avoiding overfitting, and get the final model with 0.13 MSE
WORK EXPERIENCE
Data Analyst, Avis Budget Group, Inc., Parsippany NJ August 2016 Present
Analyzed demand forecast and Revenue Management optimization model results, monitoring data integrity and results while
evaluating effectiveness, then recommended changes to be made in the optimization model
Applied decision support tools and database queries to business problems in order to access/capture appropriate data,
interpret the results, and built automated report creating process with 80% of time saved
Built web scrapper to collect car rental data and flight data with Python and stored them in the database for future analysis