DEVELOPERS TRACKInvited Talk Abstracts
Bio: Jeff Hammerbacher is the Vice President of Products and Chief Scientist at Cloudera. Jeff was an Entrepreneur in Residence at Accel Partners immediately prior to joining Cloudera. Before Accel, he conceived, built, and led the Data team at Facebook. The Data team was responsible for driving many of the applications of statistics and machine learning at Facebook, as well as building out the infrastructure to support these tasks for massive data sets. The team produced several academic papers and two open source projects: Hive, a system for offline analysis built above Hadoop, and Cassandra, a structured storage system on a P2P network. Before joining Facebook, Jeff was a quantitative analyst on Wall Street. Jeff earned his Bachelor's Degree in Mathematics from Harvard University.
Title: Socializing Big Data
Abstract: The past several years have seen some exciting developments for big data junkies. Open source projects for serving and processing massive data sets have proliferated within large, respectable companies, and large data sets have been made accessible to the general public free of charge. We'll take a look at how ecosystems develop around data sets and software projects, keeping the focus on people as system builders, algorithm designers, and data analysts.
Title: Making social networks more useful: Aardvark Social Search
Abstract: The rise of social networking tools online has an enormous potential for improving our daily lives, but most usage of social networks is for fun and entertainment. To add utility while maintaining social engagement, we looked to communication channels where people were already directly exchanging help, and then used algorithms to make the exchanges dramatically more efficient and powerful.
Algorithms can find the perfect person from within the thousands of friends-of-friends and tens of thousands of peers with common affiliations who might have the knowledge you're looking for that would otherwise be quite difficult to come by. And even though a growing amount of helpful knowledge is represented online, it remains a tiny fraction of the knowledge in peoples' heads.
For the past year we have been designing Aardvark, a new Social Search product that enables people to tap into the wisdom of their networks, conversationally, in real-time. Aardvark works over IM or Email, and is a contact that lives on your existing buddylist in your IM program. Users send Aardvark everyday questions just like they would when talking to a friend. In the background algorithms find the perfect person to answer from within their extended network. Aardvark then asks for and receives an answer within a few minutes, and passes
the answer along to the asker with a chance for follow-up in real-time.
In this talk, Nathan will discuss the challenges that faced the team in developing Aardvark. A social product needed a successful conversational interface, and the team used numerous wizard-of-oz techniques to prototype Aardvark such that it could later be automated. Users also expected a unified and extensible multi-channel product experience (web, IM, email, social network, etc), and to implement such an experience we used protocols like xmpp and legacy-network transports along with the more recently emerged Facebook Connect and Open Social platforms. To pull it all together, we incorporated user-centered design into an agile engineering process, and iteratively researched, developed and tested a variety of algorithms to find the perfect person to answer a particular question from a particular user.
Bio: Nathan Stoll is the Keeper of Insectivores at Aardvark. Nathan came from Google, where he headed Google News for three years, overseeing
its rapid expansion to more than 40 countries and growth into one of the top global news sites. In his first few years at Google, Nathan and one other engineer launched Google Suggest, helped optimize the early Adwords system and developed pricing models for Adsense. Nathan holds a BS in Computer Science and a BA in Political Science from Stanford University.