Inspiration
Japan is cool, people want to travel there. How can we help them?
What it does
Plans a vacation based on your input, and you provide feedback to improve the plan.
How we built it
Algorithmically select best vacation dates using LLM, determine optimal "solution" this way. The frontend is in javascript and allows you to interact with our pipeline. Flask and AWS are used for management, as well as S3 for user database storage.
Challenges we ran into
Llama likes to hallucinate, had to steer away from a more efficient algorithm because of this.
Accomplishments that we're proud of
The front-end turned out very well and matches the classic/arcade theme!
What we learned
How to build a giant system, there's a lot of moving parts and queries in our model.
What's next for Nihon-Go!
Fixing the llama integration so that our model can be faster, more accurate, and more unique to the user!
Log in or sign up for Devpost to join the conversation.