INTERN Resume

SUMMARY

Have two years’ intensive experience in data analytics field and be able to leverage business knowledge into modeling/machine learning process.
Solid technical knowledge of data collecting, cleaning and modeling using various tools and techniques.
Fully understand the predictive data mining process and be able to utilize data visualization skills such as D3.JS and Tableau to reach the highest satisfaction.
Excellent communication skills are facing technical and nontechnical audience, including fluency in professional English, native Mandarin, Cantonese, and basic French.
Be familiar with a variety of data mining/machine learning algorithms with a mathematical understanding such as Decision Tree, Logistic Regression, dimension reduction (PCA, SVD), regularized method (ridge and lasso), Bayesian Net and inference (EM algorithm), Support Vector Machine, and clustering .

PROFESSIONAL EXPERIENCE

INTERN

Confidential

Responsibilities:

Applying regression analysis with time errors (ARIMA) to predict the weekly conversion rate in Confidential ’ landing page, the final prediction accuracy is 87%.
Applying web analytics knowledge to use Omniture and Tealeaf to summary KPI (Key Performance Indicators) and business reports to affiliate departments, launch A\B testing experimental design on the landing page and conduct t - test.
Using CART Decision Tree to improve the model performance in predicting the mortgage closing probability from incoming calls to closing.

ProDUCT Manage

Confidential

Responsibilities:

Lead a group of 12 developers to localize QUICK. INC (Japan) financial data analytics platform and be familiar with the project management process.
Be responsible for high-frequency trading data products that consist of dealing high volume of storage (several TB per day) and deployment to mutual fund customers.

Data ANALYST

Confidential

Responsibilities:

Be responsible for financial text data extraction in corporate governance.
Performing valuation analysis on Chinese-listed banks, researched banking industry and use R for quantitative analysis on Chinese Listed Companies Top Management data to Hong Kong Polytechnic University.
Provide financial data consulting solution to customers

JAVAImplement

Confidential

Responsibilities:

Wrangling with Forest Cover Type Prediction and deal with a semi-supervised problem under the situation of unbalancing training data set and testing data set.
Implementing tentative algorithms to solve the multi-class prediction problem using multinomial logistic regression, random forest, gradient boosted trees and the EM algorithm.
Achieve top 10 on the public Leaderboard. Kaggle leaderboardUsing Java programming to modify Google Refine online document cleaning process. Add a new function to Google Refine when users can upload documents and automatically generate a copy of user’s local database. Using Java and Apache Lucene to get two million tweets from Twitter and build a retrieval model using TF-IDF indexing, tokenization and Vector Space Model for recommending information science professionals. Through this project, it is possible to leverage big data analytics using Hadoop and other relevant tools