Warmup Project

Before we start working on a sponsor’s project, it’s a good idea to practice some of our development workflow. You will work as a team on this assignment, but the teams don’t necessarily have to reflect your final team for the course.

In this project, imagine that you have an old, cranky, and vague professor named Matthew that is giving you busywork. Your mission is to figure out what exactly he wants, determine how it can be achieved, make a simple prototype design, and then implement it.

AI All the Things

Matthew spends most of his time ranting about the good old days, and complaining about how kids just use AI to drive their cars around and do their homework. Get off his lawn! However, he is at least willing to be convinced that AI is not that bad if you can impress him with an AI-based project.

What you need to do:

• Come up with an idea for a website, service, command line application, etc. that can plausibly leverage an AI model to do something helpful.
  • Once you have a solid idea, confirm with Old Man Matthew that it is indeed considered helpful. This shouldn’t just be the equivalent of “hello world”
• Design a high-level overview of the system and the components it will need. Create a diagram.
• Find a (ideally free) model API that you can use for the project.
  • Google Gemini API is free to use
  • You could set up ollama on your local machine – it has its own API and one that is compatible with OpenAI’s API
• Figure out how to interact with the model using its API
• Build the rest of your prototype
• Make a short video demonstrating how your project works

Workflow

During this project, you will need to follow our Github development workflow, so make sure to start by setting everything up for your team! You have to follow all the usual development conventions, even though this project is small.

Grading

You will be graded on the following:

• [50%] Does your project achieve the functionality you originally pitched?
• [25%] Did you follow the development workflow correctly? (Project planner, issues, pull requests, code reviews)
• [25%] A project retrospective. Each team member will turn in a document detailing what went well, problems that arose, and lessons learned.