Scale By The Bay 2018: Michelle Casbon Interview
Recording: Scale By The Bay 2018: Michelle Casbon Interview
I'm Michele Kasbah I'm a senior engineer at Google I live in San Francisco I'm working on tube flow which is an open source platform for running machine learning on kubernetes and the coolest thing about it is that it's something I really could have used before I joined Google as a data scientist at a start-up building something from scratch so it's a way to sort of shortcut all of the really nitty-gritty infrastructure details and build the machine learning application in a lot less time with a lot less effort for sure data pipelines it's very relevant last week we announced pipelines for coop low so it's definitely something that I'm interested in keeping up to speed on and kind of hearing the types of problems that people are having now my work with data science for closed data engineering we're close I've been at Google for about a year now and so I want to hear what other people are running today versus what I was running a year ago two years ago best practices I think are really important in machine learning is following the kubernetes architecture model and applying that to machine learning there's a lot of really good software developing best practices that have developed over the past few decades but they don't exactly fit with the machine learning model and so taking some of those best practices and applying those to machine learning applications that I think is really interesting because it's a way to leverage a lot of the work that we've already done in traditional software development and applying that to this sort of new paradigm in this new way of running software I'm interested to hear how people are architecting their machine learning applications without kubernetes so where are what are some other technologies that solve very specific problems and how can we sort of bring that into people and solve very specific problems that maybe we haven't looked at and we haven't tried to solve before well Scala was kind of my first my introduction to open source so it it was I was learning spark at the same time and spark was really the first open source project that I was directly involved in and it opened up a lot of doors for me it helped me think about software development in a much different way and think about how I could solve problems from a perspective that I hadn't necessarily seen before so it very much changed my ability to take a well-defined problem and solve it in a in a new way and it also provided just this this world of open source that they just hadn't been a part of before [Music] my favorite things or people Oh Martin Oh Dorothy is always a highlight it's so great to see him in person he's just been a huge example for me over the years and it's it's always great to see him and just the diversity of talks to hear people from all over the country all over the world talking about the the challenges that they're knee-deep in and how they've addressed those the biggest problems that they're facing and what they're doing to solve that that's that's always really inspiring and it gives me a sense of the direction that cue flow should be moving in to be able to address those problems what are the biggest pain points right now and what are the ways that it's as a platform we can help to make that a bit easier you