[Colloquium] Talk at the University of Chicago

[Dawn Ellis]

A student at The University of Chicago asked me to share this with you.


This week, the ACM will be hosting a talk from Affirm
(https://affirm.com/) on Machine Learning. The talk will be given by
Naftali Harris, an alum of the College and former ICPC World Finalist.

If you are planning to attend, please RSVP at http://goo.gl/PLzwX0 so we
can get enough food.

Naftali will also be doing on-campus interviews for Affirm on Friday,
November 14. If you are interested in interviewing with them, please
e-mail careers at affirm.com (you can find a list of their open positions
here: https://www.affirm.com/careers/)

TITLE: "Challenges and solutions in production machine learning"

WHEN: Thursday, November 13 @ 4:30pm

WHERE: Ryerson 276

SPEAKER:  Naftali Harris (A.B. '12), Data Scientist, Affirm
(http://affirm.com/)

SUMMARY
Broadly speaking, the goal of machine learning is to approximate some
function y = f(x) by looking at a lot of examples of (x, y) pairs.
Academic studies of machine learning tend to focus on the mathematical
and computational aspects of this function approximation problem. It has
become increasingly clear, however, that the challenges associated with
building and deploying machine learning models in production
environments are also far from trivial. At a company like Affirm, which
uses machine learning to offer unsecured loans to consumers making large
(~$1000) online purchases, it's critical to get these issues right.

In this talk, I'll identify a variety of practical machine learning
challenges, together with solutions we've found found for them. I'll
give an overview of how machine learning is used at Affirm, and describe
the problems we need to solve when building and deploying our machine
learning systems.

SPEAKER BIO
Naftali didn't graduate from high school, but fortunately UChicago was
kind enough to admit him anyway. At UChicago he took a lot of classes in
Statistics, CS, Econ, and Chemistry, refusing to take Math classes (on
principle) until his last year. He also hung out in the MacLab, played
way too aggressively at midnight soccer, did ICPC, and was lucky enough
to end up at World Finals. He graduated (for once) in 2012 after three
years with a Statistics BA, and went on to a PhD in Statistics at
Stanford. Along the way he did internships at Kaggle and Twitter, and
now he's on leave from the PhD to work full-time at Affirm as a "data
scientist". In his spare time he keeps a blog at naftaliharris.com.

-- 
*Dawn Ellis*
Administrative Coordinator,
Bookkeeper
773-834-1757
dellis at ttic.edu

TTIC
6045 S. Kenwood Ave.
Chicago, IL. 60637
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20141113/fc9608b0/attachment.htm