Shiyu

Bio

I am a M.S. Candidate from Dept. of Computer Science at Columbia University. I received my B.S. degree in Applied Math from Dongbei University of Finace and Economics(China) in 2008 - 2012, during this period I spent my senior year studying at University of Southen Canlifornia in 2011-2012 as a transfer student. Then I came to Columbia as a graduate student, during 2013-2015, I double major in computer science and statistics.

I am full-stack engineer and data scientist candidate.

Linux&C projects on topics of socket-communication & sync

• data structure design&socket api
• Linux&Sync-Http-Server with different implementations inlucidng multi-processing, multi-threading, thread pool, thread-blocking-queue

Below is a selected list of my data science projects coding in Python.

• Competition Results(Solution 1 is our group's submit)
  Spam Filter Application(Top one solution in a competition, COMS4771 Machine Lerning class Summer 2015)
• From Distance to Location(Convex Optimization Problem Sovling)
• Active Learning Illustration

And more statistical machine learning projects coding in R

Below is a selected list of projects in field of Computer Networking.

• Enhanced Chat Room with Non-Persistent TCP ConnectionCode
• Link Layer Routing Algo. Simulation Code

Below is a selected list of projects in field of Natural Language Processing.

Projects cover following areas

• Veterbi Algorithm
• Multiple Dependency Parsing
• Word Sense Disambiguation
• Machine Translation

Code

Below is a selected list of projects in field of Operating System. Due to professor's requirement, project code is not allowed to be posted in public.

Projects cover following areas

• Nonblocking I/O and I/O multiplexing
• Shell Sychronization
• System call
• File System
• Scheduling

Below is a selected list of projects in field of Information Retrieval.

In this repo, there are 3 projects which is coded together by my teammate and me. These three projects cover popular and practical topics in information retrieval field including

• query strategy design
• space vector model
• websites ranking
• association rule mining
• online data mining

Hackathon Winner Applications

FanCee --- Revolutionary video sharing experience

FanCee(A brand new experience of video watching and sharing. Win the DevFest2015 best database implementation Prize. ) Built within Python Flask framework together with YouTube API, Twilio API. All audience within a same video room share control over the video from their own PCs. Support Flying message interaction and video chat(will be expanded soon). Support phone message invitation, which sends a link to a user and automatically sets up watching environment after the user clicking it.

Instruction

1. login with your phone number and password.
2. Host a video by clicking host button. After clicking host, you will notice there is a Room No. generated for you.
3. Then you choose a video. by clicking choose a video button.You need to supply a youtube URL at this step.
4. Then you can

• give the Room No. to your friends whoever you want to share the video with.<\li>
• Invite your friend by his/her phone number. (notice this only works when your friend is currently also logged in.

SmartDonor --- won donorschoose challenge in hudson hack NYC tech intern edition hackathon.

An interactive Map embeded rich information of past 15 years' 800,000 projects across US. Support state-wide comparison and data visualization. Built with shutterstock API, Google analytics.TRY it now!

PingKong --- won semi-final of ShutterStock annual Hackathon of 2015

A semi-auto email templates builder having rich features specified for ShutterStock