The TWIML AI Podcast (formerly This Week in Machine Learning & Artificial Intelligence) Enterprise Readiness, MLOps and Lifecycle Management with Jordan Edwards - #321
Dec 2, 2019
Jordan Edwards, Principal Program Manager for MLOps at Microsoft, shares his expertise on MLOps and model lifecycle management. He discusses how Azure ML enhances collaboration between data scientists and IT teams, streamlining model deployment. Key topics include the challenges of scaling machine learning in enterprises and the importance of reproducibility and automation in evolving customer needs. Jordan also delves into the significance of a maturity model for organizations in adopting effective MLOps practices.
AI Snips
Chapters
Transcript
Episode notes
Internal vs. Enterprise AI/ML
- Internal Microsoft teams, like Bing, have extensive experience with AI/ML, influencing their tools and habits.
- These established practices may not align with how enterprises adopt AI/ML, especially regarding preferred programming languages and operating systems.
Team Structures and Challenges
- Established teams like Bing often have large, collaborative data science teams, unlike smaller, dispersed enterprise teams.
- Many enterprises struggle with managing data scientists' work, lacking code organization and hindering business utilization.
MLOps Maturity Model
- Prioritize reproducibility in your ML workflow to ensure consistency and enable automation.
- Focus on a clear path to production, including packaging, certification, and controlled rollout, before considering automated retraining.