Python,Java, SQL, R,Tensorflow, Spark, Hadoop, Linux,Bayesian models

Kun Chen

** ***** ******, *********, **, ***** 646-***-**** ******@********.***

Objective

Seeking for Data scientist/Data analyst positions. EDUCATION

Columbia University, New York, NY May 2017

Master of Science in Electrical Engineering

Relevant Coursework: Deep Learning, Machine Learning, Bayesian Models in Machine Learning, Big Data Analysis, Large Data Stream Processing, Mobile App Development Shanghai Jiaotong University, Shanghai, China May 2015 Bachelor of Science in Information Engineering

Relevant Coursework: Database, Data Structure, Operating System, C++ Programming, Probability and Statistics SKILLS

Programming: Advanced in Python.

Intermediate in Java, SQL, R, Theano, Tensorflow, Keras, Spark, Hadoop, Linux. EXPERIENCE

Real-Time Object Detection System YOLO

Final Project, Columbia University Feb 2017 - May 2017

• Built a real-time object detection system of 1,000 classes using Tensorflow and Keras.

• Simplified detection process and increased fps of system with unified detection technique to determine bounding boxes and class probabilities.

• Improved the accuracy of the system with batch normalization, anchor boxes and multi-scale training. An Artificial System to Create Artistic Style Images Final Project, Columbia University Nov 2016 - Dec 2016

• Implemented artificial system to create artistic style images from photos using deep neural networks.

• Employed VGG Network to extract content representation from content images and style representation from style images.

• Performed optimization algorithms which minimizes content and style loss to create images of high perceptual quality.

• Experimented with different parameter tuning strategies in Theano to achieve the best result. Sentiment Analysis of Large-scale Text Data Nov 2016 Class Project, Columbia University

• Built linear classifier to conduct sentiment analysis based on 1,000,000 reviews of restaurants in Pittsburgh.

• Utilized data representation using different methods and reshaped the new data in pandas.

• Applied different machine learning methods such as averaged-perceptron and Naïve Bayes to train models in scikit-learn in Python.

• Compared different representation and learning methods and selected the best method combination via cross validation.

Application of Massive Data Mining in Digital Clinical Medicine Bachelor’s Thesis, Shanghai Jiaotong University Feb 2015 - Jun 2015

• Built predictive model based on 10 years’ thyroid tumor datasets of 800 patients.

• Converted complex datasets and unstructured case records to structured datasets using Java.

• Built classifier using neural networks and implemented data mining process in R.

• Assessed the model using different criteria such as sensitivity, specificity and AUC(more than 0.9). Study of Behavioral Pattern using Data Analysis Techniques Data mining competition project, Shanghai Jiaotong University May 2015

• Analyzed campus consumption records of 3000 students and focused on the study of student behavioral patterns.

• Defined and optimized feature selection in study based on careful inspection of data.

• Clustered different individuals by predefined features and recognized different behavioral patterns.

Contact this candidate