XINYU DONG
*** ****** ******** **., ***** Brook, NY 11790.
+1-631-***-**** ac1soz@r.postjobfree.com http://xinyuqingqing.xin EDUCATION
Stony Brook University, Stony Brook, NY August 2015 - May 2017 Master in Computer Science, GPA 3.97/4.0
Tianjin University, Tianjin, China August 2011 - July 2015 Bachelor in Electronic Commerce, Distinguished Thesis Honor SKILLS AND QUALIFICATIONS
Familiar with Python Java, R, SQL.
Having experience in web application development using Java technology stack. Good at applying machine learning algorithm and statistics tools. EXPERIENCE
BMI DB Lab, Stony Brook University Stony Brook, New York Research Assitant advised by Professor Fusheng Wang August 2016 - Now
Mainly responsible for building a geometry system to nd longitude and latitude inside NY state for text address input.
Used Master Slave Based Parallelization to accelerate the program.
Combined multiple open source geology databases as data source to improve accuracy.
Used decision tree boosting classi er to enhance the searching result.
Accomplished a paper based on the system submitted to ACM Sigspatial conference.
Prediction accuracy is 96.46%, and could geocode 1 million addresses within 2000 seconds on a linux server with 44 cores.
Used Java and Python.
BioData Mining Lab, SUNY Korea Incheon, Korea
Research Assitant advised by Professor Sael Lee January 2016 - July 2016
Studied machine learning algorithms, especially in using social and temporal data to enhance recommendation algorithm.
Installed Hadoop and spark on a cluster of Ubuntu servers in laboratory. NOTABLE PROJECTS
The Nature Conservancy Fisheries Monitoring February 2017 - April 2017 Kaggle Competition Project
Aim to build a program to detect shes in pictures and to classify the shes.
Got a silver medal, ranking top 5%.
Applied VGG16 net to extract features from images, then used a shallow neural network to classify features.
Used Python, keras with theano back end.
Yelp Restaurant Photo Classi cation October 2016 - December 2016 Computer Vision Course/ Kaggle Competition Project
Built a classi er to tag restaurants with multiple labels based on photos.
Applied convolutional network(Inception-V3, VGG16) and GBDT to build a hierarchical ML ensemble classi er.
Acquired mean-F1 score 0.83206 with 54 submissions, while the original top 1s score is 0.83177 with 93 submissions.
Used Python, with scikit-learn, Ca e and tensor
ow. Research and Implementation of Recommendation System January 2015 - June 2015 Undergraduate Thesis
Studied SVD based recommendation algorithms and Implemented SVD, SVD++ and timeSVD++ algorithms.
Parallelized the SVD algorithm to improve the speed, and added movie genre attributes to improve the accuracy.
Decreased RMSE by 4% and running time by 60%.
Accomplished a paper submitted to Journal of System Engineering.
Used Python with numpy.
Java E-Commerce Web Application Development September 2013 - December 2013 Specialized Course Design Project with three classmates
Built a food E-Commerce website and other complementary systems, including e-bank, administration back to simulate E-Commerce operation process.
Used Java, Struts, hibernate and MySQL.