MLOps.community MLOps #14 Kubeflow vs MLflow with Byron Allen
May 28, 2020
Byron Allen, Senior Consultant at Servian, compares MLflow and Kubeflow, highlighting their functionalities, pros/cons, and the challenges of integrating different ML tools. They discuss managing dependencies in MLOps, the responsibility of an ML engineer in setting up MLflow, and determining the maturity level of MLOps products. The episode also explores the feasibility of using MLflow on Cloud Run and emphasizes the importance of understanding your use case in MLOps.
Chapters
Transcript
Episode notes
1 2 3 4 5 6 7 8
Introduction
00:00 • 2min
Community Slack, Guest Background, and Journey from Texas to Australia to London
01:47 • 3min
MLflow vs Kubeflow: A Comparison
04:49 • 19min
Comparing Kubeflow and MLflow in an MLOps Pipeline
24:13 • 17min
Managing Dependencies in MLOps
41:28 • 7min
Responsibility of ML engineer in setting up MLflow for experiment tracking
48:35 • 2min
Determining the Maturity Level of MLOps Products
50:37 • 2min
Running MLflow on Cloud Run and Connecting with the Speaker
53:07 • 2min