Weijie HUANG

• Mobile：(44) 07421-317-267
• Email：weijie.huang@ucl.ac.uk

Education

• Master of Science in Web Science and Big Data Analytics, UCL, 2015 — 2016
• Bachelor of Science in Computer Science, Beijing Normal University, Zhuhai, GPA 4.0/5.0, 2011 — 2015
• Government-Sponsored Exchange Student, Fu Jen Catholic University (Taiwan), GPA 4.3/5.0, 2012 — 2013
• Blog：https://koalagreener.github.io.
• Github：https://github.com/koalagreener
• Expected position：Data Scientist, Research Engineer
• Location：Any location

Projects

University College London (Oct 2015 — Feb 2016）

• NLP Assignments：The source code of three NLP projects and the reports. (Using Scala language)

Sentence Representations for Sentiment Analysis on Twitter

In this assignment, we investigate representation learning for sentiment analysis of tweets. To this end, we build our simple deep learning framework using Recurrent Neural Network and Long Short-Term Memory respectively and learn the task-specific word and sentence representations. The best accuracy result of test dataset is 77.62%.

Biomedical Event Extraction

In this assignment, sentences are mapped to structured representations of the biomedical events. We use Perceptron Algorithm and Loglinear Models, which is a set of binary feature functions to predict the biomedical events. We have implemented the feature functions in several ways, such as lexical feature,entity-based feature,syntax-based feature. Comparing to Naive Bayes Algorithm(MLE), the result of our model that trained by Perceptron Algorithm is 34.82% higher.

Language Models for processing rap corpus

In this assignment, I design and implement specific language models and apply them to the provided rap corpus. By comparing different results with different language models, I chose the best pair of the parameter to reduce Perplexity from 4216 to 336, and the result of document classification peaked at 90% and 96.67% respectively on two datasets.

Beijing Normal University, Zhuhai （Mar 2014 — May 2014）

Alibaba Big Data Competition

• Judging customers’ repurchase possibility by filtering and Slope One algorithm, and outputting result based on data (such as purchase, add to favorite and add to cart) provided by Alibaba
• Creating data filter definitions, setting up parameters and adapting algorithm to actual requirements

Experience

BNUZ Enactus (NGO) Team, Phyt-Refine Project （ Oct 2011 — Jun 2014 ）

• This project aims to Cultivate herbs in the rivers to cleanse polluted water and then refine the compound element from the herbs to produce profits and create job opportunities.
• In this three years, I mainly responsible for presenting the project during competitions, making and clipping the videos, and slides. Also, I participated in planting herbs in the watercourse and conducted an in-depth investigation in polluted rural areas and proposing practical.
• With three years of relentless efforts, Zhuhai Municipal Government moved the biggest polluting source in 2013 and invested in planting herbs at the mouth of Pearl River. As a result, water quality has been remarkably improved.
• Winning Top 10 Team (221 teams in total) Award and the Project Incubation Opportunity in 2013 Junior Achievement YESxBOP Business Plan Competition
• Winning National 2nd Prize (227 universities in total) in 2014 Enactus China National Competition

Honors and Awards

Beijing Normal University, Zhuhai

• Top Ten Graduates, 2015
• First Prize Scholarship, three times, 2012 — 2014

Greeting

I hope that we can work together in the future, and thank you for reading my resume.