Inspiration
Bad parking situation - UNCC is infamous for it's bad commuter parking situation. Live occupancy discord bot.
What it does
It helps UNCC students find the best parking garage for when they are getting onto campus.
How we built it
We used PyTorch, Pandas and NumPy for the backend and trained a model to predict what the percentage filled of the garage is when we reached. We then wrapped that in an LLM model to do decision making and give reasons as to why it chose that model. However, most of the work was done by the actual trained model to have efficiency.
Challenges we ran into
Estimating the amount of time to get to a location without paying for GMAPS API... There is no reliable method and we ended up using a distance to time module..
Accomplishments that we're proud of
The UI looks great and this is a very solid product that is usable right now by any student!!
What we learned
I learnt a lot about training models and writing backends.. Also learnt a huge amount about prompt engineering because the LLM really decided to make some questionable decisions.
What's next for Park Engage
Mobile App.
Log in or sign up for Devpost to join the conversation.