Creating reliable machine learning on unstable data, with Liftoff
Liftoff is a Californian startup helping companies to find and target the users who will like their apps and mobile services. Currently enjoying a strong international growth, Liftoff chose Paris and the Partech Shaker to strengthen its R&D in machine learning.
Correlations of the past do not always allow us to predict the future. The challenge of machine learning for Liftoff is to learn in real time from data that can suddenly change, whether for legitimate reasons or as a result of errors and bugs beyond our control.
John Whitbeck is one of the very first engineers at Liftoff. At the head of the machine learningteam, he will explain how Liftoff built a solid platform, capable of performing more than 100 million predictions per second, while adapting quickly to changes and errors in the hundreds of terabytes of data absorbed each day.
Please arrive by 9:20, as the event will start at 9:30 sharp.