Siyu Song

Machine Learning Engineer Manager 2 @ PayPal

About

An graduated M.S. student in Statistics, University of Chicago, with well-trained probability knowledge, statistical modeling and programming skills(Python, R, SQL, Matlab, C, Linux). Experienced in cutting-edge machine learning methods, especially in the neural network and deep-learning. Now working in Capitalone as a Sr. Data Analyst.

Country

United States

City

San Jose

Industry

Financial Services

Skill

Python, R, machine learning, Deep Learning

Experience

analytics technical specialist

precima

2017-5 - 2017-8 · 4 mos

Leverage data analysis and modeling skills to serve one of the largest US food and drink logistic operator; Develop a TOPX product recommend model using conditional probability and ranking algorithm.

research assistant

university of chicago

2016-9 - 2017-3 · 7 mos

* Implemented deep neural networks in Lasagne in Python and applied them to classify tens of * thousands of hand-written digit images and object images. * Implemented VAE architecture and built transformer VAE in Lasagne for generating images from original or transformed training digital sources; * Developed the variational mutual information maximization of Info-GAN in VAE.

algorithm engineering intern

douban

2016-6 - 2016-9 · 4 mos

* Processed billions of raw customer-advertisement log records in Dpark (a Spark clone) in Python; * Built and validated a Gradient Boosting model in Python for ad-recommendation, which has * been deployed on Douban.com for real-time online advertisement assignment; * Optimized the advertisement assignment strategy which improves the ad-click-rate by 30%-50% and presented the results to the department director.

risk management intern

citic securities company limited

2014-10 - 2015-1 · 4 mos

* Developed pipeline scripts for drawing real-time financial data from Bloomberg database and priced * derivatives in VBA; * Presented the analysis and conclusion regarding ABS and MBS risk in Subprime mortgage crisis to * senior executives of China Securities Regulatory Commission.

research assistant

school of physics peking university

2013-4 - 2014-6 · 1 yr 3 mos

* Publication: Plasmonic Focusing in Nanostructures. Plasmonics (2014) 9:879–886 * Produced independent graduate thesis in theoretical and experimental physics: The Numerical * Simulation and Experiment Preparation of Fano-Resonance Between Double-Body Structure * Constructed microscope device * conducted numerical simulation on commercial platform

manager, data science

capital one

data mining intern

china merchants bank

2014-7 - 2014-8 · 2 mos

* Largest non state-owned commercial bank in Shanghai, China * Forecasted daily customer service telephone traffic (200k calls/day, serving 30 mil. total clients) * using MATLAB * collaborated on a 3-member data team * Answered questions about model and asked recommendation from staffing team at daily meetings * Analyzed impact of mobile app use on telephone traffic * identified correlation between increased * mobile app usage and decreased demand for customer service telephone traffic (MATLAB) * Developed model to predict international consumption behavior for ~2 mil. clients (SAS)