This content has been archived. It may no longer be relevant
ARTIFICIAL INTELLIGENCE PRODUCT MANAGER COURSE
Learn to develop AI products that deliver business value
COURSE BUILT IN COLLABORATION WITH
Overview:
This ultimate goal of the AI Product Manager Nanodegree program is to help students learn the unique skills that define the success of a machine learning product.
Students will learn how to identify business cases that can benefit from AI technologies, and implement best practices for designing datasets and product prototypes.
Educational Objectives
A graduate of this program will be able to:
- Decide to use an unsupervised, supervised, or deep learning model when approaching a specific problem.
- Design a data annotation job to create a novel dataset.
- Build predictive models using automated machine learning tools.
- Compare the performances of learned models using suitable metrics.
- Prototype, test, and iterate on an AI product.
1
Course1
Introduction to AI in Business
AI enables innovation by automating tasks that were previously repetitive, and time-consuming.
Today, it seems like every business either depends fundamentally on the capabilities of AI, or seeks to rapidly upskill its workforce to compete in the new, AI world.
Learn the foundations of AI and machine learning, starting with the unsupervised and supervised models that are used in industry today.
Understand how to develop a clear, narrow business case for an AI application.
Learn how and when to use AI in a product based on business metrics and data availability.
2
Course 2
Create a Dataset
Training data is the currency of AI – no model will perform successfully with poor quality input data.
Learn how to develop a relevant, complete, unique and high-quality dataset.
Learn how to use Figure Eight’s data annotation platform to develop a labeled dataset for supervised learning.
Understand how to anticipate data failures and plan for longevity
Project 1: Create a Medical Image Annotation Job
Learn how to create a novel dataset, by designing a data annotation job.
In this medical image annotation project, your goal, as a product owner is to build a product that helps doctors quickly identify cases of pneumonia in children.
You’ll want to build a classification system that can help flag serious cases of pneumonia and act as a diagnostic aid for doctors.
Your main task will be to create a data labeling job using Figure Eight’s platform, https://www.figure-eight.com/ platform/.
3
Course 3
Build a Model
AI products rely upon machine learning models at their core.
Understand key fundamentals of AI models including how neural networks produce decisions and how “training” works.
Understand how training data affect the performance of a model, and how to evaluate models’ results.
Learn how transfer learning and neural architecture search make AI available to a wide variety of users.
Project 2: Build a Model with Google AutoML
In this project, get experience building models using automated ML, from data to results (no coding required).
Build your own model using Google AutoML for a medical imaging use case.
Then, implement the model with four different variants of the data in order to appreciate how the data affect the performance of the model.
4
Course 4
Measuring Impact and Updating Models
AI models are only as valuable as their impact on your business.
Learn how to measure post-deployment impact, and how to make data-informed improvements with A/B testing and versioning.
Ensure that your model continuously improves via active learning.
Understand how to avoid major failures due to unwanted bias, and how to ensure data security and compliance in different geographies.
When you finally have product market fit, learn to plan for the future and scale your product.
Project 3: Capstone Proposal
In the capstone project, you will develop a business proposal for an AI product for a use case of your choosing.
You’ll develop a business case for the product, define success metrics, scope the dataset, plan the model development, and build a postdeployment monitoring plan.
Reviewers will evaluate your proposal for rigor and completeness.