The Joe Reis Show 5 Minute Friday - The Quality Paradox
6 snips
Nov 8, 2024 The speaker reflects on their experiences at the MLOps World Conference, emphasizing practical applications of machine learning. They dive into the pitfalls of poor data modeling and advocate for coherent structures to boost productivity. A compelling discussion on prioritizing quality over speed reveals how high standards prevent technical debt. Personal updates provide insight into the speaker's professional journey, highlighting upcoming engagements and shifts in focus.
AI Snips
Chapters
Transcript
Episode notes
Messy Data Migration
- Joe Reis describes a friend's data migration project.
- It involved converting SSIS jobs to DBT, resulting in a complex, tangled DAG resembling a rectangle.
Prioritize Data Modeling
- Create a comprehensive data model upfront.
- This improves quality, productivity, user happiness, and pride in your work.
The Point of Data Ops
- Joe Reis argues that the data ops and ML ops movements often miss the point.
- The true goal is increased quality, which reduces defects and errors, not just monitoring.