Botong Hu (Jeremy)
http://botonghu****.x**host.com
**** ******* ******, *** #*** 201-***-****
Hoboken, NJ 07030 ************@*****.***
Business Intelligence Analyst/Data Scientist
A top-performing computer scientist, applied mathematician and recent
graduate with a Master of Science in Electrical Engineering received from
Stevens Institute of Technology. 3 years of experience using statistical
models and undertaking database analysis. Ability to confidently extract
and query very large data sets using Java, Python, SQL, and other
programming languages while successively refining applicable parameters to
accurately express data trends or anomalies.
Education/Training & Certifications
[pic]
Master of Science in Electrical Engineering, Data Engineering, Stevens
Institute of Technology, Hoboken, NJ (2014)
Relevant Coursework: Data Structure and Algorithm Machine Learning and
Data Mining Pattern Recognition and Classification Database Management
System
Bachelor of Communication Engineering, Beijing University of Chemical
Technology, Beijing, China (2013)
Certified National Psychological Consultant, Level 3 Certificate (2012)
Computer Program Expertise
[pic]
MATLAB/R/Java/C/Python/HTML/PHP Scaling/Performance/Schedul Windows/Mac OS/Linux
ing MapReduce
MySQL/ PostgreSQL ETL Techniques/ Data Machine Learning Algorithm
Data Mining/ Data Modeling and Cleansing
Optimization Hadoop
Relevant Professional Experience
[pic]
Lenovo, Inc. (Beijing), Beijing, China 2012
Product Assistant, Consumer Product Marketing Department
. Essential mobile product assistant tasked with consumer channel survey
and data analysis, CDMA mobile products planning and consumer electronic
channel strategy planning.
. Supported development, execution, forecast and management of action plans
for new product introduction and product line extensions.
. Provided technical services with product information, responded to
customer questions, requests and assisted in responding to formal
complaints or suggestions regarding products and product lines.
Academic Projects
[pic]
Stevens Institute of Technology, Hoboken, NJ 2013 - 2014
A Hybrid ARIMA and Support Vector Machines Model in Stock Price Forecasting
(Fall 2014)
. Implemented a support vector machine model (SVM) for predicting Tesla
Stock open value ranging from Jun 29th 2010 to Nov 31st 2014 based on
optimizing parameters in kernel function.
. Built an ARIMA model; developed linear regression of Tesla stock open
value.
. Created a model by combining two models; obtained a smaller MSE in
testing data compared to SVM and ARIMA models, working primarily in
MATLAB.
Outcome
. The MSE of hybrid model, ARIMA,SVM is 0.000084, 0.036, 0.00086
respectively; new hybrid model combining the advantage of SVM and
ARIMA increased performance on forecasting.
Improved Predictive Model by Combining Neural Network and Genetic Algorithm
(Spring 2013)
. Pre-processed data with principle component analysis in MATLAB; reduced
data dimensions in order to avoid over-fitting.
. Built Back Propagation neural network with gradient descent algorithm;
predicted NBA players' free throw percentage in playoffs.
. Created a better way of obtaining optimal weights matrix in neural
network based on genetic algorithm; significantly reduced the Mean Square
Error in testing set compared to BP neural network; worked primarily in
MATLAB and R.
Outcome
. Mean square error of improved model is 5.3704*10-5; MSE of neural
network model is 1.8776*10-4; shows improvement of smaller error than
back propagation neural network.
Recommendation System Methodology in Yahoo Music (Spring 2013)
. Built hash table for storing hierarchy relationships between tracks,
albums, artists and genres using Java.
. Implemented Yahoo music recommendation system with large dataset using
java based on neighborhood method to recommend 3 songs with higher scores
to each user given users' rating history.
Outcomes
. Compared to the real result, the accuracy reached 78%.
. Published thesis in 2012 was 64%; results show an increase of 21.8%.
Beijing University of Chemical Technology, Beijing, China
May 2012
Project Leader of the Miniaturization Design of Bluetooth Antenna
Organized group brainstorming sessions; assigned tasks to team members
after discussion.
Designed plans to optimize and simulate Bluetooth antenna.
Implemented new structures with PIFA's strength and finished the
emulation of the antenna,
Laid out the antenna and led the group to finish research report and
summary.