Explore Git-based CI/CD for machine learning in this one-hour webinar. Learn how to implement continuous delivery of ML models to production using Git-based pipelines with hosted training and model serving environments. Discover techniques for automating workflows, reviewing models, storing versioned artifacts, and running CI/CD for ML projects. Gain insights into enabling controlled collaboration across ML teams using Git review processes and implementing MLOps solutions with open-source tools and hosted ML platforms. Watch a live demonstration showcasing the practical application of these concepts, covering topics such as deploying serving functions, building topologies, and creating pipeline workflows.
Overview
Syllabus
Introduction
The problem
The siloed approach
The production approach
What do we need to do
Problem 1 Silos
Operational Pipeline
Transformation
Serverless
Four Big Pillars
Feature Store
Operations Store
Realtime Pipeline
CICD Automation
ML Project
Pipeline Exports
ML Run
Training
Demo
Deploy serving function
Build topology
Mock server
Pipeline notebook
Pipeline workflow
Taught by
Open Data Science
