CTO at Bantr
San Francisco Bay Area
CTO at Bantr
San Francisco Bay Area
I'm currently working on building out Bantr and Circle of Friends, and helping Team Rankings however I can.
My background is in math / stats / analytics / algorithms; I love to solve technical problems, optimize things, and tinker (in lots of ways).
Statistical modeling and analytics, viral marketing, a/b testing, decision trees, network analysis, fraud/risk, facial recognition, semantic matching, interactive search, sports ratings and predictions, investigative reporting, multivariate statistics, making stuff scale, Java, Perl, PHP, SQL, AJAX.
(Internet industry)
June 2007 — Present (1 year 5 months)
Circle of Friends, our primary product, allows people to build small communities of friends who have common interests and backgrounds. It is one of the largest applications on Facebook (over 8 million installs), and is also available on MySpace, Bebo, and hi5.
We are working on building out the next set of great targeted features for our users and their circles.
(Internet industry)
January 2000 — Present (8 years 10 months)
Team Rankings offers up a wide range of evaluative and predictive tools and analyses for sports fans. Team Rankings and its users annihilate their competition in NCAA Tournament pools: our algorithmic March Madness predictions outperformed 32 out of 33 NCAA basketball experts, number crunchers in a Wall Street Journal competition.
(Privately Held; 51-200 employees; Internet industry)
February 2004 — January 2007 (3 years)
Ran LinkedIn's analytics team. Responsible for:
-Designing Analytics Prototyping Engine, to test new personalized, analytics-driven content
-Defining and tracking LinkedIn's important business metrics
-Building intelligent algorithms for standardizing data
-Building statistical models to match users with targeted jobs, other similar users, portions of the site they may find interesting, and other relevant content
-Defining a framework for testing the effect of different interface elements on LinkedIn's metrics
-Understanding and working to optimize LinkedIn's viral growth
(Public Company; 1001-5000 employees; ebay; Internet industry)
December 2000 — August 2004 (3 years 9 months)
I was responsible for much of PayPal's back-end fraud modeling. I designed and coded a program which allowed me and my colleagues to easily build predictive mathematical models to track down fraud.
Using my software, I built about twenty statistical models to assess the risk of PayPal's users, and my colleagues used it to build another 30-40 models. The models were then implemented on the site's back end (by one of us and/or engineering) to find fraudulent transactions before money from them left the system.
(Educational Institution; 10,001 or more employees; Higher Education industry)
September 2001 — June 2003 (1 year 10 months)
Helped students build a women's squash team at Stanford. Coached practice three times a week, and took team to Nationals and other matches.
Much to the benefit of the team, I've been succeeded by a far more accomplished coach: Mark Talbott, former world #1 player, and former national championship coach at Yale.
(Privately Held; 11-50 employees; Internet industry)
February 2000 — December 2000 (11 months)
Designed and implemented systems for classifying a web site's users based on responses to multiple choice questions. Built up a Bayesian clustering algorithm, to allow for straightforward but meaningful marketing segmentation.
(Public Company; 10,001 or more employees; hpq; Internet industry)
June 1999 — December 1999 (7 months)
Built a tool to analyze web site behavior.
B.S., Mathematical and Computational Science, September 1996 — March 2000
Graduated with Honors in Mathematical and Computational Science
running, squash, cycling, soccer, cooking