Organization: University of Maryland
Personal Biography: Jay Pujara graduated from George Washington High School in Charleston, WV, taking every science class offered, and was an NYSC delegate in 2000. These experiences helped him when he attended superhero school, learning among other things, how to turn sand into computer chips, computer chips into robots, program those robots to find treasure and rescue people, and teach computers to read minds. In more conventional terms, he attended Carnegie Mellon University and got degrees in Computer Science, Cognitive Science, and Computer Engineering with minors in Robotics, Philosophy and Math with a Masters Thesis on Feature Selection in fMRI brain scans. Jay subsequently worked for Yahoo! Inc, discovering new ways to reject billions of spam messages each day at the nation’s largest webmail provider. Encountering deep mysteries in machine learning, Jay is now pursing a PhD at the University of Maryland, working on making machine learning work in the real world where computational resources and annotated data are often limited. In his spare time, Jay enjoys hiking, reading and playing board games.
Lecture Topic: The Art of Computer Science
Mr. Pujara has long felt Computer Science is misunderstood – it’s not just about writing code or the province of all-powerful hackers as Hollywood would have you believe. It’s about pancakes. And solving problems using computational tools. But also pancakes. This lecture will discuss how to think like a computer scientist when it comes to data and pancakes, and popular approaches to solving problems computationally. He will illustrate these ideas with his experiences from fighting spam on the Yahoo! Mail team and research he has done on sentiment prediction on Twitter.
Directed Study Topic: A Brief (yet helpful) Guide to Machine Learning
Machine learning is a discipline that seeks to make sense of all of the information that surrounds us. This directed study will provide an introduction to the major concepts and algorithmic approaches used in Machine Learning. There will be a number of practical examples from Bayesian classification for spam detection to topic models to automatically categorize documents, with a focus on hands-on computation of learning models.