Educom

An AI powered student matchmaking and community app.

Comment

Inspiration

We have all experienced the struggle of not understanding certain topics and not really having the connections to work alongside us. We all know that University is a lot more independent than previous education stages, because of this we wanted to build something to help students connect and communicate leveraging AI to be able to match more efficiently and get tailored assistance on specified topics.

What it does

Our app allows students to post requests and, using our AI pipeline, connect with other students with similar requests. Features include: Anonymity function which hides users' email and name in order to protect privacy. Custom rooms for greater control over users' learning environment. All built with an easy to use interface with communities and rooms easily accessible.

How we built it

For the front-end, we used react native, using Codex and Claude to assist us. The backend is coded within Javascript, alongside typescript, and the database is hosted on MongoDB atlas and is connected via node.js - running on aws.

Challenges we ran into

We ran into a few challenges with MongoDB Atlas. These included problems with connecting to a cluster in order to connect our front-end to real data in the back-end. There was a bit of a learning curve for it, but with the help of mentors, AI and a bit of resilience, we were able to get the cluster connected and working.

We also had problems using AI to help debug as sometimes the AI's would mix up logic and change certain functions. This was fixed by looking through the code thoroughly and making sure everything connects in the right places with the right logic.

Accomplishments that we're proud of

The ability to connect users via rooms to each other is a great accomplishment, thus marking our main idea complete. This is similar to a game like system where random players are matched to each other.

What we learned

There were unexpected errors with users manipulating the tab before leaving the rooms -> where if they didn't leave before closing the tab, they wouldn't be counted for leaving. This made us more aware of certain issues with not updating the database and falsely looking like the user count was inflated.

What's next for Educom

We have planned a roadmap for progress towards Educom if we were to continue it, this would involve: Implementing more sophisticated AI algorithms, this means that AI overviews and an AI chat assistant might be involved.

Enhancing Auto-Matching accuracy, this is because theres always outliers that have very vague results. Account management, so that users can track their ratings and that they can make friends/connections between other users. Maintaining old chats, this allows users to be able to refer to old results instead of constantly looking for new results to the chat. Leaderboards - a fun way to encourage users to help and ask questions in the room systems.

Share this project:

Updates