Content

• Google Drive folder (Week 10):
• Zoom link

Fantasy sports involves creating a team of individual players in a sport and then competing against another team of individual players selected by an opponent. If the players on your team exceed the statistical performance of your opponent's, you win. For example, in the matchup below, the team called Hustle defeated the team called Get Buckets because the aggregate statistics of the individual players on Hustle outperformed the aggregate statistics of the individual players on Get Buckets in 5 out of the 9 categories (3-point shots made, points, rebounds, steals, and blocks) while losing in 4 out of the 9 categories (field-goal percentage, free-throw percentage, assists, and turnovers).

Before playing each other, teams set a lineup of active players while leaving their inactive players on the bench. A team manager must decide ahead of time the set of players from their team that has the best chance of winning a matchup. LLMs can be helpful in suggesting different strategies for making this line-up decision.

In this exercise, a spreadsheet containing the projected weekly statistics for the active players for both Hustle and Get Buckets is given as well as the projected weekly statistics of the currently inactive players on Hustle. Using this spreadsheet, we'll use an LLM to help examine strategies for making potential lineup changes. Visit the course folder for the week, then create a project or a NotebookLM notebook that includes the spreadsheet (Fantasy_Basketball_Projections). Ask an LLM to examine the 3 sheets in it to find lineups for your team, your opponent's team and your bench. Assuming you are the manager for Hustle, prompt for advice on strategies for making lineup decisions such as:

What lineup changes might I make to maximize my chances of just winning?

What lineup changes might I make to maximize my chances of winning the most 
categories?

What lineup changes might I make to maximize my chances of losing?

• Add a slide with the decision-making advice given

In this exercise, you will experiment with the use of LLMs to help you decide on a purchase of a car. As in prior exercises, we will utilize a prompt in which the LLM is configured to grill you on all aspects of the purchase decision, before presenting you with a narrow list of options that you can then choose from to make a final decision. Some things that the bot should help you consider might include:

• New or used
• Lease or buy
• Buy outright or finance
• Gas, electric, or hybrid
• Car, SUV, van
• All-wheel drive
• Cost of ownership (gas, insurance, maintenance)

Configure a custom chatbot using a prompt similar to the one below.

Role: You are an expert car purchasing consultant that will help me in deciding 
which vehicle I should purchase or lease. Your goal is to interview me relentlessly 
about every aspect of this decision until we reach a shared understanding of what I 
am interested in choosing from. Walk down each branch of the decision tree, 
collecting as many of my constraints and preferences as possible. 

When we have reached an understanding, output a table of 3 potential choices I can 
make and how each one lines up with the shared understanding we've created.

Then, interact with the chatbot to produce the result.

• Add a slide that includes the results of the chat.

Personal financial management requires informed decision-making based on the goals of the individual, their risk tolerance, and the variety of options that are available for them to choose from. In this exercise, we'll utilize an LLM to decide on what to do with $10,000 that you have just inherited. In this exercise, you will utilize a custom chatbot to decide which investment vehicle to put the money into such as a retirement account, a certificate of deposit, a stock exchange-traded fund (ETF), etc.

Role: You are an expert financial investment planner that will help me in deciding 
what to do with $10,000 I have just inherited. Your goal is to interview me 
relentlessly about every aspect of this decision until we reach a shared 
understanding of what I am interested in choosing from. Walk down each branch of 
the decision tree, collecting as many of my goals and preferences as possible. 

When we have reached an understanding, output a table of 3 potential investments I 
can make and how each one lines up with the shared understanding we've created.

Then, interact with the chatbot to produce the result.

• Add a slide that includes the results of the chat.

For homework, you will do a synthesis of multiple techniques you have learned in this class to build projects you may find interesting or personally useful to you in your future endeavors. Examples might include building an LLM application using a coding agent that produces social media posts, creating a business and marketing plan for a product, or creating a data-driven decision-making and visualization application.

Screencast

Upon completing your project, via a narrated screencast of no longer than 10 minutes, you will perform a demonstration and walk-through of the prompts and artifacts produced. Ensure that the video camera is turned on initially in your screencast.

Rubric

We will be using the following rubric to evaluate your homework.

Submission

Upload your completed screencast on MediaSpace. Ensure that it is published as "Unlisted". Then, in Canvas, submit the URL that your unlisted screencast on MediaSpace is located.