What's this project for?

This project was designed for 1st Intern Open Hackday Competition (July 29~30, 2011). This competition was hosted at Linkedin Headquarters. 170 interns from various tech companies in the bay area participated and competed with the best and brightest in Silicon Valley. Each team got 24 hours to work on any project they wanted. The winners were picked by three judges: Kevin Scott (VP Engineering at LinkedIn), Omar Hamoui (Partner at Churn Labs, Founder of AdMob) and James Gosling (Creator of the Java programming language).

Our project Next Generation Social Recruiting got the 2nd place in this competition.

Update (07/12):

• The prototype has been added to the Linkedin Product Team's roadmap!
• A related (but not much) paper by us has been published in KDD'12: Magnet Community Identification on Social Networks. A very interesting paper to rank comapnies based on talent flows.

Overview

Talents are priceless, and the connections between talents are even more valuable. Companies always want to hire the right person, since hiring (s)he means hiring a group of talented people through her/his connections. We designed the next generation social recruiting intelligence system. It is capable of predicting the potential job changes for each individual and the potential influences on their professional networks through an easy-to-use user interactive interface. With the state-of-the-art data mining and visualization technologies, we believe this system can greatly improve the recruiting process as well as benefit massive job seekers like us.

Contributions:

• Huge Business Opportunities
• Innovative Visualization
• Great Research Potential

Team Members

We are three PhD students from the Next Generation Data Mining and Social Computing (NGDS) LAB at University of Illinois at Chicago led by Professor Philip S. Yu.

Our team members listed alphabetically by last name:

• Guan Wang: intern at Linkedin, Mountain View, CA
• Xiaoxiao Shi: intern at AT&T Labs, San Francisco, CA
• Yuchen Zhao: intern at eBay Resesarch Labs, San Jose, CA

Technology

Our system was implemented using pure Java. It is cross-platform and can be run as a standalone application and web application (Java applet). We used the following open source tools/libraries to build our application:

• Processing (Visualization Language)
• Gephi (Graph Layout)
• Weka (Data Mining & Machine Learning Modeling)

Demos

Due to the data availability and privacy issues, the demo is currently offline. The following videos demonstrate the main functions we implemented. Please Click play to watch. (Depending on your network connections, the videos may take a few minutes to load.)

FAQ

Detailed behind-the-scenes ideas, Predicting algorithms, Extracted features, ...

TBA

External Links

Official Hackathon Website about LinkedOut

Official Blog about the Hackathon

Official Engineering Blog about the Hackathon

Youtube Video