Building MLOps POCs and Sandbox Environments Using k3s and Argo

Explore the challenges corporations face when evaluating technologies for data pipeline automation and MLOps workflows in this 25-minute conference talk from KubeCon + CloudNativeCon North America 2021. Learn how open source technologies can reduce implementation time and infrastructure costs for MLOps systems. Discover the advantages of using k3s as a lightweight Kubernetes for POCs and the lessons learned when implementing MLOps with Scikit Learn for ML Models and Apache Spark for ETL processes. Understand how Argo can replace or complement pipelines designed with Apache Airflow and prepare companies for Edge Computing. Watch a short demo showcasing a basic MLOps workflow implementation using k3s and Argo, demonstrating their interaction in deploying ML models ready for use. Cover topics including introduction, basic concepts, differences between DevOps and envelopes, Argo UI, workflows, testing, deployment, rollback, pipelines, Python code, and inference.