Inspiration
We are inspired by dating apps which can match people with similar interests and preferences. Why can't we do that for finding friends too? As international students at university, we found it challenging to make new friends outside of our daily classes. But here, we're not trying to reinvent the wheel. We shift the interaction pattern so knowing new people can be more fun and less boring!
What it does
The app is targeted at unimelb students, matching potential friends with similar interests and hobbies, which is done using an AI model that computes similarity scores based on user-inputted preferences. People who have been matched can chat on the app. The app also has an added incentive of a points system, in which users can collect points through meeting matched friends in real-life. Points collected could potentially be used for shop discounts throughout the unimelb campus. Interesting features also include avatar generation using user photo uploaded upon registration.
How we built it
We built it using Java and Flask. We're not using GPT or another Gen-AI platform for the AI; instead, we built it from scratch using OpenCV libraries. We also code in native using Java and Android Studio to ensure the user have the best user experiences
Challenges we ran into
Several integrations are really hard, especially using virtual device. Also finding suitable model was one of the reasons keeping us awake
Accomplishments that we're proud of
WE ARE NOT GPT WRAPPER YEAY
What we learned
Don't be a deadliner LOL
What's next for Friendy
Built With
- ai
- android-studio
- api
- csv
- flask
- java
- machine-learning
- no-sql
- opencv
- python
Log in or sign up for Devpost to join the conversation.