The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Live from TWIMLcon! You're not Facebook. Architecting MLOps for B2B Use Cases with Jacopo Tagliabue - #596
Oct 24, 2022
Jacopo Tagliabue, a former AI director at Coveo and member of South Park Commons, brings a fresh perspective on MLOps tailored for B2B companies. He discusses the practical challenges these smaller firms face compared to larger tech giants, emphasizing data quality, automation, and resource allocation. Jacopo also explores the balance between custom and off-the-shelf ML tools, the importance of end-to-end ML professionals, and how community engagement can enhance hiring. His insights make machine learning operations more accessible for diverse practitioners.
AI Snips
Chapters
Transcript
Episode notes
Early Data Logging
- Jacopo Tagliabue's startup logged detailed user data early on, including rough geolocation.
- Although not immediately useful, this data later allowed analysis of geographical factors.
Buy, Don't Build
- Prioritize buying existing tools and platforms over building custom solutions.
- Focus development efforts on core business logic and value-adding ML models.
B2B Team Composition
- B2B ML teams often don't require ultra-optimizers focused on fractional gains.
- Prioritize engineers who can rapidly iterate and deliver end-to-end solutions.