Alabaro

Build a ML Workflow For Scones Unlimited On Amazon SageMaker

Project Introduction

In this project, you will build an image classification model capable of detecting the type of vehicle used by delivery drivers. This model is essential for routing drivers to the correct loading bays and orders, optimizing operations for our client, Scones Unlimited.

Background

Image classifiers have a broad range of applications, from autonomous vehicles and augmented reality to eCommerce platforms and diagnostic medicine. In this context, our model can help Scones Unlimited in several ways, including detecting people and vehicles in video feeds, enhancing social media engagement, and identifying defects in their products.

Project Scope

In this project, I assume the role of a Machine Learning Engineer at Scones Unlimited, and so my responsibilities include creating a scalable and robust model. This model should not only classify vehicles efficiently but also be adaptable to varying demand while maintaining performance standards. Ensuring that safeguards are in place to monitor and control model drift or degraded performance is a critical part of this project. This project was carried out using Amazon SageMaker

Project Steps Overview

This project is divided into four key steps:

Step 1: Data Staging

• This step involves preparing and staging the data for training, ensuring it is clean and properly formatted.

Step 2: Model Training and Deployment

• You will build the image classification model using AWS SageMaker. After training, the model will be deployed for real-time inference.

Step 3: Lambdas and Step Function Workflow

• Implement AWS Lambda functions to build supporting services and use AWS Step Functions to create a seamless workflow for your model and services.

Step 4: Testing and Evaluation

• Thoroughly test and evaluate the model's performance to ensure it meets the required criteria.