As machine learning matures, we need to find better ways to integrate data science with software development. In this talk for DivOps, a conference about the future of DevOps, DVC data scientist Elle O'Brien discusses how CI/CD can adapt to machine learning. This is MLOps, explained with fuzzy animals.
Last week, DVC was part of DivOps, a fully remote conference led by women in DevOps. DevOps, to the newly anointed, is a discipline bringing together strong software engineering practices with speedy development cycles. As machine learning is finding its way into just about every area of research and development, we're going to need to come up with some conventions and tools for integrating machine learning and big data with software development. This growing field is called MLOps.
I gave a lightning talk about how we'll have to rethink our software development practices in the age of machine learning. It's got a focus on CI/CD, a way of structuring workflows that we think can streamline exchanges between data scientists and software engineers. And, it's got fuzzy animals. Check it out here:
If you liked this, you'll also want to check out the next talk in the DivOps playlist by Anna Petrovicheva, Founder and CEO of Xperience AI. Anna's talk goes deeper into developing best practices for software engineering with deep learning.
All the talks from DivOps are available online now, so please check out the YouTube channel. And stay tuned on our blog for more CI/CD discussions coming soon…