Inspiration
- "Hey, want to add me on Instagram?"
- "Sorry, I don’t trust the Zucc. How about X?"
- "Nah, I don’t do X after the incident. What about Viber?"
- "What's a Viber? What about Myspace?"
- "That’s funny. No. What about Bluesky?"
- "Nah, I don’t even know what that is. BeReal?"
- "Missed the notification. LinkedIn?"
- "Seriously? Snapchat?"
- "Deleted that in 2016. WhatsApp?"
- "Too much drama."
As avid NETWORKERS, we have all scrambled to get our phones out trying to get everybody at the networking social to type in their LinkedIn/Instagram/WhatsApp etc. WHAT IF there was a near-instant way to receive all the socials of every new person you meet?
What it does
Introducing MaiCard, a minimal and easily accessible web application to supercharge your networking prowess. Curate your own social business cards and embed them in QR code format. Save other profiles with a single scan. Find your most compatible peers with our ML powered compatibility score.
How we built it
We built the web application with a React frontend and a FastApi backend for a seamless experience similar to Google Wallet when navigating between routes and requesting data from our Postgres DB. Our DB contained 3 tables: users, profiles, saved(profiles). Embeddings are generated for profiles, while the saved table links profiles together upon scanning to generate compatibility scores.
Challenges we ran into
Strict typing of TS slowed us down and Docker broke on us a lot.
Accomplishments that we're proud of
Fully fledged frontend, backend API endpoints, integrated DB, and embedding model trained.
What we learned
What's next for MaiCard
Login, better UI/UX, more powerful model.
Built With
- docker
- fastapi
- huggingface
- pgvector
- postgresql
- python
- react
- restapis
- supabase
Log in or sign up for Devpost to join the conversation.