MLOps as a Team - Raphaël Hoogvliets
55 minutes Posted Nov 8, 2024 at 6:00 pm.
DataTalks.Club intro
Career journey and transition into MLOps
Dutch agriculture and its challenges
The concept of "technical debt" in MLOps
Trade-offs in MLOps: moving fast vs. doing things right
Building teams and the role of coordination in MLOps
Key roles in an MLOps team: evangelists and tech translators
Role of the MLOps team in an organization
How MLOps teams assist product teams
Getting feedback and creating buy-in from data scientists
The importance of addressing pain points in MLOps
Best practices and tools for standardizing MLOps processes
Value of data versioning and reproducibility
When to start thinking about data versioning
Importance of data science experience for MLOps
Skill mix needed in MLOps teams
Building a diverse MLOps team
Best practices for implementing MLOps in new teams
Starting with CI/CD in MLOps
Key components for a complete MLOps setup
Role of package registries in MLOps
Using Docker vs. packages in MLOps
Examples of MLOps success and failure stories
What MLOps is in simple terms
The complexity of achieving easy deployment, monitoring, and maintenance
0:00
55:36
Show notes
We talked about:27 :56 Standardizing practices in MLOpsJoin our Slack: https://datatalks .club/slack.html
This podcast uses Anchor (from Spotify) for stats, tracking, and dynamic audio. Learn more on this podcast's About tab. Dynamic Audio may cause this episode's timestamps to be slightly off.