Published Oct 4, 2020

MLOps Coffee Sessions #11: Analyzing “Continuous Delivery and Automation Pipelines in ML" // Part 3

Demetrios Brinkmann and Davida Ponte delve into the essential components of MLOps such as data management, feature stores, and model performance monitoring, while highlighting the impactful role of orchestration tools and automation in evolving machine learning pipelines from manual processes to sophisticated automated workflows.

Episode Highlights

Topics covered

Data Management
01
Model Performance Monitoring
02
Orchestration Tools
03
MLOps Maturity Levels
04

MLOps Coffee Sessions #11: Analyzing “Continuous Delivery and Automation Pipelines in ML" // Part 3

Topics covered

Popular Clips

Monitoring Model Performance

Data Science Essentials

Serving Machine Learning

Data Architecture Insights

Event-Driven Automation

Experimental Recording Insights

Model Training Insights

Coffee Breaks Matter

Feature Store Insights

Automating Training Pipelines

Feature Store Insights

Data Preparation Insights

Orchestrated Validation

Orchestration Tools Explained

Episode Highlights

Data Management

Model Performance Monitoring

Orchestration Tools

MLOps Maturity Levels

MLOps Coffee Sessions #11: Analyzing “Continuous Delivery and Automation Pipelines in ML" // Part 3

Topics covered

Popular Clips

Episode Highlights

Data ManagementThe discussion shifts to the intricacies of data preparation and the essential role of feature stores in machine learning pipelines. These components are vital for ensuring model accuracy and consistency across various environments.

Data Management

Model Performance MonitoringAutomatic model monitoring and adaptation are essential for maintaining ML model performance. Davida Ponte and Demetrios Brinkmann discuss tools and strategies to address concept drift and ensure models remain effective in dynamic environments.

Model Performance Monitoring

Orchestration ToolsThe discussion shifts to orchestration tools and their role in automating MLOps workflows. Davida Ponte and Demetrios Brinkmann explore how these tools facilitate continuous training, validation, and event-driven automation, enhancing the efficiency of machine learning processes.

Orchestration Tools

MLOps Maturity Levels