Leverage Cloud Computing Benefits with Serverless Multi-Cloud ML

Build a powerful, multi-cloud ML model without writing complex algorithms. This talk shows you how to connect the dots between serverless services.

#1about 10 minutes

Defining AI, machine learning, and data science

Key concepts like computer science, data science, artificial intelligence, and machine learning are defined and differentiated.

#2about 3 minutes

Understanding the machine learning development lifecycle

The typical machine learning cycle involves fetching data, cleaning it, training a model, evaluating performance, and deploying to production.

#3about 3 minutes

Defining the problem of juggling pattern detection

Initial research reveals existing models are inadequate, leading to the decision to use computer vision and object detection for the problem.

#4about 3 minutes

Manually labeling data with Azure Custom Vision

The initial pre-processing step involves manually labeling juggling objects in images using Azure Custom Vision, a time-consuming and unscalable process.

#5about 3 minutes

Why data cleaning is critical for model performance

Using raw, user-generated content without cleaning leads to poor model performance, highlighting the necessity of filtering data before training.

#6about 4 minutes

Automating the data pipeline with multi-cloud services

A multi-cloud pipeline using AWS, Azure, and Google Cloud services automates data collection, cleaning, and preparation for model training.

#7about 3 minutes

Training, evaluating, and debugging the ML model

The model is trained and evaluated using both Azure and Google Cloud platforms, revealing some humorous misclassifications along the way.

#8about 2 minutes

Deploying the machine learning model with Docker

The trained model is exported as a Docker container, enabling easy and consistent deployment across local environments and multiple cloud providers.

#9about 4 minutes

The role of cloud services in democratizing AI

Cloud platforms democratize technology by providing managed services that reduce the required expertise and time to build and deploy complex applications.

#10about 4 minutes