Inspiration
Buying a first car is scary. We want people to feel confident in their decision and make the most informed decision they can.
What it does
Utilizes a LLM powered search engine that returns SQL queries that populate the page with tailored cars for each user.
How we built it
Built it by scraping the Toyota website and getting data. Then we utilized a Postgresql database to store everything. It takes in a prompt and then uses a LLM to decipher how to format that prompt into a SQL query. Calls the SQL query and then populates the page.
Challenges we ran into
Getting data was difficult. GCP database got hacked 3 times and we had to start over.
Accomplishments that we're proud of
Proud of getting our #1 priority finished.
What we learned
Learned a lot about GCP. Also learned a lot about SQL and the different ways that it can be utilized. Using LLMs to convert to SQL queries was very interesting.
What's next for Toyota Prophet
We would like to have filters that the user already inputs that can maybe narrow their search down even more. This and we want continued searches to narrow down the user's dream cars.
Log in or sign up for Devpost to join the conversation.