First Last
[email protected]
+86 130 0000 0000
Beijing, China
Work Experience
Miaozhen System, Machine Learning Engineer, Beijing 2019.02 - 2019.05
●Worked at advertisement traffic anti-cheating algorithm team, processed 2
billions of data daily, produced half million extra revenue for the company
● Decrypt the MD5 encrypted logs and improved the current algorithms
accuracy from 60% to 63%
● Implemented a rule based algorithm which is counting the data as invalid
traffic (IVT) with same cookies but different IDFA, IMEI, MAC using java,
Hadoop, Map Reduce, Distributed Computing on Linux Server
Microsoft, Azure, Software Engineer, Seattle 2018.02 - 2018.04
● Used EDA to analyze the malicious software data, selected features and
predict using XGBoost
● Familiarized the Internal Offices Tools such as Skype, OneDrive, OneNote,
Teams
● Read Azure official documents and get known the server deployments and
services provided
Education
US Public School
Master of Science, Computer Science, Statistics
● Relevant Course: Machine Learning, Artificial Intelligence, Data Mining,
Pattern Recognition, Database Management, Big Data Analytics and Text Mining
, Data Structure and Algorithm, Computational Geometry, Regression,
Interpretation of Data
China Regular University
Bachelor of Science, Finance, Ranking: 1/200
● C/C++, Calculus I & II, Linear Algebra, Probability, Java - Object
Oriented Programing, Discrete Mathematics, Operating System,
Computer Networks, Partial Derivative Equation
Skills
● Professional: Java, Python, C/C++, Hadoop, Linux, R, Github, Git, Qt,
Sublime, Pycharm, Eclipse
● Familiar: SQL, Scala, Spark, Objective-C, AWS, Apache Hive, Mongodb,
MySQL, Excel
● Machine Learning : Scikit-Learn, Tensorflow, Caffee, Keras, Theano, Torch
, mxnet, DeepLearning4J
Projects
Survival Analysis of Titanic
● Did Exploratory Data Analysis to data including Histogram, Pie Chart, and
Probabilistic Statistics
● Used Feature Engineering to fix the missing data, transformed the data
format to fit the model
● Implemented Logistics Regression, Random Forest, GBDT and XGBoost,
achieved 90% accuracy
A Movie Recommendation System Based on Collaborative Filtering
● Achieved a mean squared error rate of 0.944
● Collected data from MovieLens and predicted users’ rating on certain
movies
● Used Pearson Correlation Similarity Measure to cluster users as
collaborative filtering
● Applied both K-means and category to cluster movies and compared results
of different clustering
Image and Digits Classification based on Naive Bayes and Perceptron
Classifier
● Extracted features from handwritten digits and face pictures and convert
it into zero-one matrix
● Estimated parameters such as prior and posteriors from training data
● Used Naive Bayes and Laplace smoothing achieved accuracy rate of 84%
● Implemented Perceptron algorithm to classify digits and face pictures,
achieved 97% accuracy
