-Osgood Vogler: Celestica is a supply chain solutions company. We're a global company and we work with many of the world's biggest brands that includes leading companies in aerospace and defense, health tech, capital equipment, and enterprise communications. We're a team which has a lot of engineers, a lot of supply chain professionals, a lot of people who are very at home and excel. But now many of these people are empowered with a platform that allows them to scale what they're doing at a much more intelligent way and begin to really look at data as an asset. We're talking easily terabyte scale data, what we were looking for with Hunch was a giant shrink rate and what we found was that you could achieve a 30,000 times compression on that data without losing fidelity that was important for analyzing quality. There's a deep neural network learning process happening that results in that massive compression ratio but it's really if I look at it from a business perspective it's a way to make it accessible to solve certain kinds of big data problems where previously there would have been a barrier to being able to start projects at that scale. In our early adoption of this tool set we were able to show the potential to automate 97% of a manual test result interpretation on one of the important operations in our manufacturing process. From a response time perspective, what we saw with Hunch was quite remarkable. What previously would have been a 30 minute response time on a query is down to milliseconds and that allows us to perform iterative analytics in a much different way. There's no more penalty associated with running the model. That took a project that might have been 3-6 months long and compressed it down to 2 weeks.
