Scale By The Bay 2019: Thomas Gerber Interview
Recording: Scale By The Bay 2019: Thomas Gerber Interview
[Music] I'm Thomas Gerber the work for Salesforce and I work on the island platform which is an AI layer of the Salesforce platform and I own the ancient derelict that's used to feed or provide data to ml systems within Salesforce [Music] my talk is about lessons we learn building a data Lake what's interesting about the data space is it's still very much in flux compared to other complex spaces for example mlo AI which I feel are much better defined and commoditized but the data space is still extremely vibrant which means the century that we learn a lot still make lots of mistakes and so very happy to share with the community mistakes we've made and what we are all not missing mistakes but in precision that we've made and that we're correcting in the hope that that might be helpful for them in their own context I'm very happy to do that for me specifically you know I'm talking about how we change how we are adding a new abstraction in a data leak so it's a different abstraction that we think is better so I'm explaining why that's a very good question the AI and mash running in general are getting more and more prevalent in the industry and think of it this way Salesforce all the companies almost you Salesforce right it's a very good solid system of record business process automation the next step for Salesforce is to generate a lot of useful pieces of data for its consumer right so that's the next big evolution Salesforce is currently transitioning into and my work helps with that transition because we want to generate as many predictions for the customers as possible because we think it's going to help them be more efficient and be better with their customers [Music] yes I'm looking for the serverless panel tonight that's with my former boss so that's gonna be very exciting [Music] yes I think a couple of things I think we are seeing more and more database that are transaction log first I think that's a pretty exciting change which allows which allows the consumer of that database to be able to see any point in time what else as I said the data space is still pretty much in flux and there are so many things that we're learning on a daily basis on how to build scalable simple systems which is extremely hard and I think like it's it's a very good journey to be on we're learning a lot [Music] so it's interesting I think like scale by the bay started pretty much of the Scala conference almost and that's by the way how I discovered the first time I talked about house column was the right tool to write big data jobs I think that one of the suitors and I would have is to diversify continue to diversify because I think so ecosystem for for data transform for data movement is growing much faster than just Kara so I think that would be my suggestion [Music]