K-Mon

Inspiration

The problem space we were hoping to address with this project was the divide between the resources available for rich versus poor students. According to the Hechinger Reportlink here, elementary school students can gain an extra five months of education from the personalized tutoring services.

What it does

Our project is a fun way for students to interact with a LLM that can help curate questions on specific topics for students while immersing them in a fun adventure based on the universe of their choice.

How we built it

We built out two initial subject paths, where we randomize subjects, here based on the Ontario 4th grade curriculum here. Where for the math path specifically, we used a javascript code block to generate the questions that are fed through to chat-gpt which builds out a plot line placed on the user name inputs and chosen exploration world that students input in the beginning.

Challenges we ran into

LLMs like GPT have complications when it comes to processing math questions so prompt engineering came in handy here to be able to get the correct math questions that made sense within the story line. However, the functionality and ease of Voiceflow made it easier to edit changes.

Accomplishments that we're proud of

We're happy to have combined GPT with other tools to be able to structure questions in a story environment, and have been able to collaborate effectively in a short period of time, whilst troubleshooting a lot of prompt engineering questions along the way.

What we learned

Our previous exposure to prompt engineering before this project was low, so learning this skill over the course of the hackathon was invaluable in this day and age of AI.

What's next for K-Mon

We would like to develop this project more to include increasing levels of difficulty for each of the levels, but also provide developing and encouraging story lines (with more simple questions) for learners that are struggling with the content.

Built With

• gpt
• javascript
• voiceflow