Machine Learning for Real-time Bidding on Spark

Real-time bidding has seen 70% yearly growth, and is set to grow even faster as more premium inventory becomes available through private exchanges. I'll talk about the technological developments at VideoAmp - a demand side platform for video that uses machine learning and integrates multiple data sources to optimize bids in real-time. And to improve targeting accuracy, our machine learning system combines data from users' online behavior, and panel data such as TV viewership and offline purchases.

To meet the challenges of dealing with high volume and high frequency of data, I'll discuss a big data architecture using Spark. With Spark, we have seen major efficiencies in running machine learning algorithms in batch mode, and also developed real-time analytics on the same platform.

Speaker's Bio
Deb Ray (http://www.linkedin.com/in/debray) is VP of Data Products at VideoAmp, a DSP that optimizes buying across TV, online video and mobile through a unified platform. Deb completed his PhD in Machine Learning and Behavioral Economics at Caltech. At Caltech, he founded Pasadena Labs, an adtech company that optimized search marketing and successfully automated management of thousands of campaigns for market aggregators of small businesses. Deb has also developed computer vision based products at Microsoft Research and machine learning algorithms for financial (high-frequency) trading.

Pizza will be provided at around 6:15pm and talk will start at 7:00pm sharp!