Data Graphic Designer

Quan Yuan

San Jose, CA ***** *****@*****.*** 734-***-**** GitHub

EDUCATION

University of

Michigan

MS in Entrepreneurship

**** - **** *** *****, **

GPA: 3.71 (Top 10% of

Business graduate students)

Tsinghua University

BS in Industrial

Engineering

BS in Economics

2010 - 2014 Beijing, China

Outstanding Practicum (top

1 out of 60)

GPA: 3.84

COURSEWORK

• Machine Learning

• Data Manipulation

• Explorative Data

Analysis

• Information Retrieval

• Data Structure &

Algorithm Design

• Operations Research

• Probabilistic User

Behavior Modeling

• Database Design

• Experimental Design

• Applied Statistics &

Probability

• Management

Information Systems

• Simulation

• Econometrics

• Game Theory

SKILLS

Languages

Python • Java • R •

C/C++ • MATLAB •

MySQL • PHP •

HTML/CSS • JavaScript

Tools

Numpy • Scipy • Pandas •

Scikit-learn • Theano •

NLTK • Spark •

Matplotlib • CakePHP

WORK EXPERIENCE

Data Scientist (Summer Intern)

June 2013 to August 2013 Foxconn Technology Group - Yantai, China

• Defined Task Complexities Metrics; built Linear Regression models with Lasso or Ridge Regularization to predict Ramping-up Capacity of Production Lines; estimated coefficient intervals via Bootstrapping, achieving a 16.2% increase in accuracy.

• Applied Genetic Algorithm to select the best match between work stations and workers regarding task complexities; implemented the Online Forecast System and a simple personnel management system. Product Manager (Summer Intern)

May 2015 to August 2015 YONO Health Inc. - Sunnyvale, CA

• Created the low fidelity prototype for the mobile app using and managed the graphic designer and the programmer on mobile app development.

• Managed the hardware development progress of the outsourced design firm.

• Created usability test protocol under FDA guidance; also built function testing plan and marketing testing survey.

ACADEMIC PROJECTS

Statistical Analysis on Taxis in New York City

School of Information, University of Michigan - Ann Arbor, MI

• Cleaned the NYC taxis operations data in 2013 (about 300 million rows) using PySpark and visualized taxis’ distribution based on the location data converted from Bing Map API.

• Predicted Tip Amount by stacked models of GLM, MARS, P-splines Regression and Gradient Boosting Trees with L1 post processing, tuned models by AIC and/or GCV and evaluated and visualized effects of variables.

• Implemented ISLE (Importance Sampled Learning Ensembles) based on Sklearn. Statistical Quality Analysis of Tea Polyphenols (TP) in Tea Drinks Data Analytics for Quality Excellence (DATE) Lab, Tsinghua University – Beijing, China

• Collected and cleaned the raw profile data (spectrum curves), reduced its dimensions from 1751 to 30 using discrete wavelet analysis and denoised the data using soft thresholding method.

• Applied SVM model to identify different brands of tea drinks based on the cleaned data; tuned model parameters by cross validation.

• Evaluated the most important features by minimizing Generalized Classification Error using Genetic Algorithm.

Sentiment Analysis on Tweets

School of Information, University of Michigan - Ann Arbor, MI

• Cleaned 1.6 million tweets (tokenization and Named Entities replacement) in dataset Sentiment140 using MRJob.

• Implemented statistical language model for tweets classification; Tuned the model parameters by cross validation and reached the accurate rate of 94% on 16K sample data.

• In another dataset, Twitter US Airline, built a similar model and identified words with top predictive contributions.

• Implemented Discrete Hidden Markov Model for POS tagging and sped up its performance by Cython. Otto Group Product Classification

• Trained and tuned multiple Machine Learning models (KNN, Logistic Regression, Ensembles, SVM, etc.) for stacking. Trained a Deep Belief Network with the output of these ML models as input.

• Implemented a Deep Neural Networks Module on top of Theano.

Contact this candidate