AI Product Manager Certification earned by MoniGarr

• monigarr

MoniGarr earned the AI Product Manager Certification on July 22nd 2021 from the Udacity Nano Degree program.

AI Product Manager Syllabus:

https://www.udacity.com/course/ai-product-manager-nanodegree–nd088

  • Intro to AI in Business:
    • Intro to AI and Machine Learning
    • Using AI and Machine Learning in Business
  • Creating a Dataset
  • Build a Model
    • Training & Evaluating a Model
    • Project: Build a Model
  • Measuring Impact & Updating Models
    • Measuring Business Impact & Mitigating Bias
    • Case Study: Video Annotation
    • Project: Capstone Proposal. Request Access to Review MoniGarr’s Project Work regarding “Open-Source solution to train custom models for translation, summarization, language modeling and other text generation tasks for an eastern dialect of Kanien’kéha (Akwesasne dialect).

CAPSTONE PROPOSAL FEEDBACK

Congratulations Monica, you’ve passed your final project in this nanodegree 🏆!

Thanks for your note about your draft, I think you’ve underestimated the quality of your first draft. This is really detailed and by Udacity world standard (well I reviewed students around the world), your draft is a final proposal. The section that you’ve left blank (ML metric), you’ve actually answer in the business metrics section, with just one small detail missing – I have provided the guidance in that section. Put simply, you did well 💯

What a beautiful project submission – this is truly an amazing work. Thanks for all the visuals, it helps to grasp what you intend to do quickly. What a joy reading your Project Proposal. Loved your project. Please do continue to learn and push yourself.

This is the final project for this nanodegree, nevertheless, I’d encourage you to continue reading / dig deeper into AI product management. Here’s a few good sites for you:…

Your project is really interesting, let’s hope you’re able to get this going in real life. And one last thing, I’d like to leave this rather interesting way of shaping AI & ML project, hope this will take you forward further.

 

CAPSTONE PROJECT SUCCESS CRITERIA:

Success Criteria Specifications

Are all required files submitted?

The submission includes a Proposal file (a pdf) and images of the user interfaces of the proposed product.

Is the proposal complete?

Every section of the proposal file has been completed.

Business Goal

Success Criteria Specifications

Does the proposal describe a business problem?

The proposal includes a short description of the problem the product seeks to address. The problem is described in business terms; for example, the problem might be the need to increase customer satisfaction or drive repeat customership. What is the business benefit of your proposed solution?

Does the proposal make a business case for the product?

The proposal contains an argument in favor of the product that derives from impact on revenue, market share, and/or other drivers of business success.

Does the proposal describe a narrow task that AI/ML could help solve?

The proposal should describe what the AI/ML model will actually do.

Success Metrics

Success Criteria Specifications

Do the success metrics relate to the business goal(s)?

The success metrics measure how well the product achieves the business goal(s).

Data

Success Criteria Specifications

Does the proposal describe how the data will be acquired and discuss issues that may arise in acquisition?

The proposal should discuss the following considerations: buying vs. collecting data; privacy/personally identifying information (PII)/sensitivity issues; cost; ongoing data vs. one-off data dump (which would need to be refreshed).

Does the proposal describe the size of the data? Does the proposal describe potential biases in the data?

The proposal should include an estimate of the size of the data. The proposal should describe the categories/types of data that will be under/overrepresented in the dataset(s).

Does the proposal justify the choice of data labels?

The proposal should explain the proposed labeling scheme, and why the chosen labeling scheme was chosen. What are the strengths and weaknesses of such a labeling scheme?

Model

Success Criteria Specifications

Does the proposal describe how the model will be built?

The proposal should include a description of how the model will be built, and should discuss considerations such as the likelihood an external platform will satisfy the specific use case, the need for certain controls on the model, data sensitivity/security, and willingness to give an external platform access to the data.

Does the proposal describe the planned use of ML metrics to measure the performance of the model?

The proposal should describe how ML metrics such as accuracy, precision, recall, F1 score, etc., will be used to assess the performance of the model.

Minimum Viable Product (MVP)

Success Criteria Specifications

Does the proposal include sketches of main user interfaces of the product?

The proposal should include a few sketches of main user interfaces of the product.

Does the proposal describe the “types” of users who will use the product?

The proposal should describe a few prototypical users and their use cases.

Does the proposal lay out a general plan for building and launching the product?

The proposal should have a general pre-launch and post-launch plan.

Post-MVP-Deployment

Success Criteria Specifications

Does the proposal describe strategies to improve the product in the long term?

The proposal should describe a plan for testing, versioning, and active learning (learning from new data).

Does the proposal include a plan for monitoring bias?

The proposal should mention a plan for mitigating unwanted biases.

Suggestions to Make Your Project Stand Out

  1. Including data to support your arguments will make your proposal stronger.
  2. Detailed drawings/mock-ups of your product, or prototypes/demonstrations of the ML/AI functionality will certainly make your proposal shine!