Mike Del Balso (@mikedelbalso, CEO at @TectonAI) talks about lessons learned from Uber’s Michelangelo ML platform, enabling DevOps for ML data, and how Tecton enables features for data models.  

SHOW: 477

SHOW SPONSOR LINKS:

• Learn more about Fauna: https://www.fauna.com/serverless
• Try FaunaDB for Free: https://dashboard.fauna.com/accounts/register
• CloudAcademy -Build hands-on technical skills. Get measurable results. 
• Get 50% of the monthly price of CloudAcademy by using code CLOUDCAST
• Datadog Security Monitoring Homepage - Modern Monitoring and Analytics
• Try Datadog yourself by starting a free, 14-day trial today. Listeners of this podcast will also receive a free Datadog T-shirt.

CLOUD NEWS OF THE WEEK - http://bit.ly/cloudcast-cnotw

PodCTL Podcast is Back (Enterprise Kubernetes) - http://podctl.com

SHOW NOTES:

• Tecton homepage
• Tecton emerges from stealth with veterans from Uber
• Michelangelo: Uber’s Machine Learning Platform
• Tecton: The Data Platform for Machine Learning (blog)
• “Why We Need DevOps for Machine Learning Data” (blog)

Topic 1 - Welcome to the show. It’s always exciting to talk to new companies. You were doing some pretty interesting things at Uber prior to starting Tecton, so tell us a little bit about that experience and then what motivated you to start Tecton? 

Topic 2 - There are lots of Data/AI/ML tools and platforms out there. Tecton talks about “great models need great features”. Give us a high-level overview of the Tecton platform and the perspective you bring to solving complex business problems.

Topic 3 - After reading the papers on the Uber Michelangelo platform, it’s clear that today’s interactions aren’t a bunch of individual “decisions”, but layers of decisions made on ever-changing data (the UberEATS example). Why does business need a new approach to how they interact with data? 

Topic 4 - When I think about earlier approaches for companies to “harness data for analytics”, there was always the problem of data silos. Do you find that companies need to organize themselves different, not just organize their data, to be able to overcome those silo challenges? Does it take a much more product-centric approach vs. the traditional “analyst” approach?

Topic 5 - Every new company and platform needs to find product-market fit. What do you see as early “fits” for the Tecton platform? 

Topic 6 - How much data-science expertise does a company need today to be able to leverage Tecton, and how much does the platform lower the barrier to entry? 

FEEDBACK?

• Email: show at thecloudcast dot net
• Twitter: @thecloudcastnet