MLOps concepts will be taught with a real-world case study i.e. estimating daily energy generated from the sun(solar energy) using Abuja city. It will be implemented using Machine Learning and Deep Learning FrameWork.

Objective

The aim of this course is to teach AI/Data Scientist/ML Engineers to understand MLOps, and also how to deploy models across different resources and different cloud platforms.

At the end of the session, you will be able to have:

• Knowledge of MLOps implementation.
• Know how to deploy machine learning and deep learning models across different cloud infrastructures.
• Understand the automation of model deployments.

• Fundamental knowledge of how to develop and train machine learning and deep learning models.
• Some intermediate knowledge of python programming is necessary. And some knowledge of flask or fast API development will be an added advantage.
• Basic understanding of cloud computing: cloud storage(s3, azure blob, google cloud bucket), cloud virtual machine instances(ec2, data store ).

A glance at ML Life Cycle

• Challenges facing MLOps

Introduction to MLOps

• What is MLOps?
• Why the need for MLOps?
• Where & when do we adopt MLOps
• Components of MLOps
• Introduction to APIs
• Challenges and the need for APIs in MLOps

Containers for  ML Deployment

• Introduction to Docker
• Introduction to kubernetes
• Deploy machine learning models using docker
• Deployment of containers on kubernetes(EKS, GKE, etc)
• An introduction to automating ML deployment workflow

Leveraging Cloud Computing for MLOps

• Deploying machine learning model through AWS
• Deploying deep learning model though google cloud
• Train and deploy ML model through Azure Auto ML
• Deploy model via Fastapi, Streamlit, Heroku

Monitoring and Automation

• Overview of Monitoring
• System infrastructure monitoring
• Data pipeline monitoring
• Monitor and evaluate model performance
• Maintenance guide for model updating

An introduction to CI/CD for automated model deployment