Michael avatar

Michael Zhou

Computer Science M.S.(General) at University of Southern California
Research Assistant at Marshall School of Business, University of Southern California, under supervision of Yu Ting Forster Wang and T.J Wong
Email: Michael Zhou
Github: Georgy Zhou
Linkedin: Qijia Zhou
Phone: +1(323)318-4110

Brief Curriculum Vitae

I am currently a first-year computer science master student in Viterbi School of Engineering, University of Southern California. I received my Bachelor's degree of engineering in computer science from Zhejiang University in June 2016. My interests mainly lie in Machine Learning and Text Mining. I was an undergraduate research assistant at the State Key Lab of Computer Aided Design & Computer Graphics from September 2014 to June 2016, supervised by Prof. Hongxin Zhang researching on data table visualization, knowledge base system and visual reasoning. I also worked for AZTechX Co., Ltd as a software engineering intern from October 2015 to January 2016. Now I am engaged in text mining over financial as well as international affairs reports under supervision of Yu Ting Forster Wang and T.J. Wong. You can download my full CV through here


  • Software Engineering, tools and Infrastructure Intern at Google Inc
  • Research Assitant at Marshall School of Business, USC
  • Previous Software Engineering Intern at AZTechX Co.,Ltd
  • Previous Software Engineering Intern at Qingmedia Co.,Ltd
  • Previous Co-Founder of ELINC (an on-campus entrepreurial team)
  • Previous Technology Minister at Google Camp, Zhejiang University
  • Previous Research Assistant at State Key Lab of Computer Aidede Design & Computer Graphics
  • Previous Member of "Wang-Daheng" Fellowship, Zhejiang University & Chinese Academy of Science
  • Previous President of Computer Science Student Union, Zhejiang University
  • Previous Chief Editor and Sales Director of Future Dream 20th and 21st edition
  • Previous Mathematics & Physical tutor


I am reading JVM details and developing tools in Google these months:)

Selected Projects

2016 ByteCup, Extracurricular, Python, Team leader in a team of two
  • Developed a service that intelligently distributes online questions to the most suitable experts to answer.
  • Tried support vector machine, collabrative filtering and Nerual Network. Used collabrative filtering to reach the final result
  • Ranked top 40 over more than 500 teams
  • Detail information is available here
Live Call Quality Rating System, Extracurricular, C/C++/Matlab, Developer in a team of two
  • Explored and exploited live call quality rating system that converts, analyzes and rates a live video stream
  • The most difficult part is how to give user-oriented and real-time analysis based on incomplete as well as unstable H.264 stream. Creatively solved the problem by applying motion detection, optimizing matrix operation and frame detection
  • Won the second-class Prize(Prized by Huawei) and Ranked 3rd in 2016 Huawei Cup Innovation Technology Competition
Babytoy Android Tablet Game, Extracurricular, Java, Individual Developer
  • Interact with game device with bluetooth to make 4-year-old to 8-year-old child to learn about insects.
  • The project is finished in half a month and can be accessed through here.
  • It was intended for my girlfriend's undergraduate capstone design:)
ShotVis, Data Table Visualization System, Extracurricular, Python, Individual Developer
  • Established a knowledge base with ProbCog(Alchemy based) and independently developed a system to recommend(Markov Logic Network) and implement(ECharts) optimal visualization for a specific table
  • Challenging and brand new topic which means coming up with original idea without any reference, setting up own dataset(no existing dataset), analysising and extracting table features, reasoning and visualizing
  • Demo here
TrojanHacks 2.0 FaceRank, Extracurricular, Java, Individual Developer
  • Devloped an Android app that ranks a facial image based on Microsoft Face API and ranking model
  • Designed and developed for USC TrojanHacks Hackathon.
  • Project respository here
Pascal Compiler, Course Work, C/C++, Developer in a team of two
  • Implemented the whole compiler purely by C/C++, with use of Lex & Yacc
  • The compiler can parse the grammer of Pascal and compile Pascal codes to a executable file
  • The total number of code lines is over 3000
ERPC, Estimated Revenue Per Click on ads, Internship Projects, Python/R, Developer in a team of three
  • Implemented a real-time service to predict the revenue of a specific ads keyword based on random forest
  • Trained an ads keyword classifier and added the semantic-based feature to ERPC. Precision increased by 8%
  • Avoided unnecessary loss significantly by filtering some bad ads keywords out
Ads Marketing Tool, Internship Projects, Java/HTML/SQL/MongoDB, Developer in a team of two
  • Collaborated in a team of two to develop an ads marketing tool which implements concurrent online management, market data monitoring, job scheduling, database integration and user interface
  • Aimed to develop a unified tool to meet kinds of ads operation demands
  • This tool is now used by tens of operating stuff in AZTechX Co.,Ltd
Github User Ranking System, Extracurricular, Python, Developer in a team of two
  • Ranked Github users with Heterogeneous Information Networks(HINs) constructed on 500000+ pieces of Github user and repository data
  • Compared with PageRank, HINs showed both higher precision and recall rate over Github community data
  • Crawled the data from Github by using github API and web crawler
3D Modeling Software, Course Work, C/C++, Developer in a team of two
  • Developed software with OpenGL that supports model file(self-defined) import/export, zooming in/out, wandering, collision detection, dragging object, texture mapping, basic elements constructing such as cubes, spheres and cones
  • Most challenging point is to integrate OpenGL and Qt into a unified tool that correctly responses to cursor operation
  • Got special grades on this project
New York Times Mining, Research Project, Python, Individual Developer
  • Data collection and data cleaning: extract PDF-format article; crawl financial report/article
  • Applied topic model(LDA) on reports/articles and established internal connection
  • Used TABARI and automatic script to extract events from these articles and reports
Stand-alone RPG Game, Course Work, JavaScript/C#, Lead in a team of four
  • Player plays a character in the game to find necessary objects to get out of a locked house. The whole game is based on conversation prompt and player inferences,
  • It was finished by using Unity3D