Skip to content

Case study · Nanodegree

Udacity AI for Business

Three projects from Udacity's AI for Business Nanodegree: a medical image labeling job, a pneumonia classifier in Google AutoML and a model proposal.

My role
Solo
Team
Solo
Timeline
Sep – Nov 2021
Context
Udacity Nanodegree
Platform
Appen and Google Cloud AutoML
Udacity AI for Business: screenshot 1
Fig. 01 · Udacity AI for Business: screenshot 1

01 Overview

I completed three projects, each covering one stage of an AI product.

  • DataDesigned a medical image annotation job on Appen for a chest X-ray dataset from Kaggle
  • ModelTrained a pneumonia classifier on chest X-rays with Google AutoML and wrote its evaluation report
  • ProposalWrote a capstone proposal for a vehicle identification model

02 Problem

Each project covered one stage of an AI product: the data and its labels, the model, then the business case.

03 My role

I did all three projects on my own.

AI and data
Google CloudAppen
Web
HTMLCSS

04 Key decisions

Platforms over hand-built models

What we chose

Appen for labeling and Google AutoML for training

Why

The program was about the decisions around a model, not its code.

Trade-off

AutoML hides the model's details.

05 Result

Completed in November 2021; the evaluation report and the proposal are in the repo.

A chest X-ray labeled with the lungs, heart, spine, ribcage and diaphragm, from the Appen annotation project
Fig. 02 · A chest X-ray labeled with the lungs, heart, spine, ribcage and diaphragm, from the Appen annotation project

What I learned

These projects showed me how to turn modern technology into real value for people.

With thanks to

Next project: Trip app wireframe

Mobile flight search, wireframe

Questions

Want to Go Deeper on Udacity AI for Business? I’m Happy to Walk You Through It.

Email me