Productionizing Real-time Serving with MLflow

Overview

Explore how to elevate MLflow serving to production-grade status in this 28-minute conference talk from Databricks. Learn about deploying machine learning models as REST API endpoints and advancing them to containerized production environments. Discover techniques for implementing custom middlewares, monitoring, logging, and performance optimization for high-scale applications. Gain insights into Yotpo's approach to making MLflow serving production-ready, covering topics such as continuous delivery, request transformation, exporting metrics, deployment strategies, and monitoring best practices. Delve into optimizations and control mechanisms to enhance your MLflow serving capabilities for real-world, high-performance scenarios.

Syllabus

Introduction
Continuous Delivery
MLflow Serving
Request Transformation
Exporting Metrics
Deployment
Monitoring
Optimizations
Control

Taught by

Databricks

Reviews

Start your review of Productionizing Real-time Serving with MLflow

Taught by

MLflow Model Serving - From Training to Production

Productionizing ML Models Using MLflow Model Serving

Automated Monitoring in Production for Continuous Model Improvements

Never Stop Learning.