Dive into a comprehensive 2.5-hour video tutorial on building MLOps platforms using Databricks, AutoML, SKLearn, and AWS. Learn to start an MLOps project, set up CI/CD, and invoke ML library code. Explore end-to-end MLOps with Databricks to AWS containers, spin up Databricks clusters, and transition from Pandas to Spark. Create a fake news classifier using Kaggle and AutoML, register models, and set up inference endpoints. Utilize Github CodeSpaces and FastAPI to serve MLFlow models with Swagger documentation. Discover Iguazio's feature store capabilities and leverage AWS Cloud9 for developing containerized ML models. Finally, deploy your containerized model using AWS App Runner. Access source code and additional resources for further learning in cloud computing, data engineering, and MLOps.
MLOps Platforms From Zero - Databricks, MLFlow-MLRun-SKLearn
Overview
Syllabus
Intro
Starting an MLOps project
Setup CI/CD
Invoke ML Library Code
End to End MLOps with Databricks to AWS Containers Diagram
Spinning up Databricks Cluster
Doing Pandas to Spark
Creating Fake News Classifier using Kaggle and AutoML
Creating Databricks AutoML Experiment
Viewing Databricks AutoML Experiment notebook
Registering models with Databricks
Setting up Inference endpoint with the Databricks platform
Using Github CodeSpaces to serve out downloaded Databricks model with MLFlow
Using FastAPI to serve Swagger documentation of MLFlow model
Feature Store Capabilities of Iguazio
Using AWS Cloud9 to develop containerized ML Models
Using AWS App Runner to serve out containerized model
Taught by
Pragmatic AI Labs
