#120 Applying ML Learnings - Especially About Drift - To Data Mesh - Interview w/ Elena Samuylova

Sign up for Data Mesh Understanding's free roundtable and introduction programs here: https://landing.datameshunderstanding.com/Please Rate and Review us on your podcast app of choice!If you want to be a guest or give feedback (suggestions for topics, comments, etc.), please see hereEpisode list and links to all available episode transcripts here.Provided as a free resource by Data Mesh Understanding / Scott Hirleman. Get in touch with Scott on LinkedIn if you want to chat data mesh.Transcript for this episode (link) provided by Starburst. See their Data Mesh Summit recordings here and their great data mesh resource center here. You can download their Data Mesh for Dummies (info gated) here.Elena's LinkedIn: https://www.linkedin.com/in/elenasamuylova/Evidently AI on GitHub: https://github.com/evidentlyai/evidentlyEvidently AI Blog: https://evidentlyai.com/blogIn this episode, Scott interviewed Elena Samuylova, Co-Founder and CEO at the ML model monitoring company - and open source project - Evidently AI. This write-up is quite a bit different from other recent episode write-ups. Scott has added a lot of color on not just what was said but how it could apply to data and analytics work, especially for data mesh. Some key takeaways/thoughts this time specifically from Scott's point of view:A good rule of software that applies to ML and data, especially mesh data products: "If you build it, it will break." Set yourself up to react to that.Maintenance may not be "sexy" but it's probably the most crucial aspect of ML and data in general. It's very easy to create a data asset and move on. But doing the work to maintain is really treating things like a product.ML models are inherently expected to degrade. When they degrade - for a number of reasons - they must be retrained or replaced. Similarly, on the mesh data product side, we need to think about monitoring for degradation to figure out if they are still valuable or how to increase value.Data drift - changes in the information input into your model, e.g. a new prospect base - can cause a model to not perform well, especially against this new segment of prospects. That data drift detection could actually be a very...

Om Podcasten

Interviews with data mesh practitioners, deep dives/how-tos, anti-patterns, panels, chats (not debates) with skeptics, "mesh musings", and so much more. Host Scott Hirleman (founder of the Data Mesh Learning Community) shares his learnings - and those of the broader data community - from over a year of deep diving into data mesh. Each episode contains a BLUF - bottom line, up front - so you can quickly absorb a few key takeaways and also decide if an episode will be useful to you - nothing worse than listening for 20+ minutes before figuring out if a podcast episode is going to be interesting and/or incremental ;) Hoping to provide quality transcripts in the future - if you want to help, please reach out! Data Mesh Radio is also looking for guests to share their experience with data mesh! Even if that experience is 'I am confused, let's chat about' some specific topic. Yes, that could be you! You can check out our guest and feedback FAQ, including how to submit your name to be a guest and how to submit feedback - including anonymously if you want - here: https://docs.google.com/document/d/1dDdb1mEhmcYqx3xYAvPuM1FZMuGiCszyY9x8X250KuQ/edit?usp=sharing Data Mesh Radio is committed to diversity and inclusion. This includes in our guests and guest hosts. If you are part of a minoritized group, please see this as an open invitation to being a guest, so please hit the link above. If you are looking for additional useful information on data mesh, we recommend the community resources from Data Mesh Learning. All are vendor independent. https://datameshlearning.com/community/ You should also follow Zhamak Dehghani (founder of the data mesh concept); she posts a lot of great things on LinkedIn and has a wonderful data mesh book through O'Reilly. Plus, she's just a nice person: https://www.linkedin.com/in/zhamak-dehghani/detail/recent-activity/shares/ Data Mesh Radio is provided as a free community resource by DataStax. If you need a database that is easy to scale - read: serverless - but also easy to develop for - many APIs including gRPC, REST, JSON, GraphQL, etc. all of which are OSS under the Stargate project - check out DataStax's AstraDB service :) Built on Apache Cassandra, AstraDB is very performant and oh yeah, is also multi-region/multi-cloud so you can focus on scaling your company, not your database. There's a free forever tier for poking around/home projects and you can also use code DAAP500 for a $500 free credit (apply under payment options): https://www.datastax.com/products/datastax-astra?utm_source=DataMeshRadio