Professional connections, without the BS.
The antithesis of LinkedIn – real-time, genuine professional connections via sleek UI, effective AI-powered smart matching, and live voice/video calls.
LinkedUp was born out of the frustration with platforms like LinkedIn – cluttered interfaces, bot-filled interactions, and an overall unproductive networking environment. We reimagined professional connections by building an app that emphasizes:
- Real Human Interaction: Instant voice and video calls powered by Stream.
- Seamless Onboarding: Robust authentication and onboarding with Clerk.
- Smart Matching: An ML model that pairs professionals with like-minded peers.
- Sleek Design: A minimalist UI that focuses on productivity and genuine networking.
Engage in live voice and video calls that bypass endless text exchanges.
Our ML-driven system pairs you with professionals who share your interests and ambitions.
A clean, distraction-free environment that lets you focus on building meaningful connections.
A quick overview of the core technologies powering this project.
| Category | Technologies |
|---|---|
| Monorepo 🚀 |  All-in-one monorepo management |
| Front-End ✨ |      Hosted on Vercel |
| Authentication 🔒 |  |
| Real-Time Communication 📡 |  Voice/Video Calls |
| Backend & Database 💾 |        Hosted on Render |
| ML Matching Engine 🤖 |   Vector-based similarity (dot products, cosine/euclidean distance) + semantic analysis using OpenAI |
.
├── apps
│ ├── web # Next.js + Tailwind + Shadcn UI front-end
│ └── backend # FastAPI, Python, XGBoost pipeline, etc.
├── packages
│ ├── shared # Shared libraries/config
│ └── ui # Reusable UI components (React, TS, Shadcn)
├── turborepo.json # Turborepo config
└── ...
- Front-End deployed on Vercel
- Back-End deployed on Render
Before you begin, ensure you have met the following requirements:
- A modern web browser
- Node.js installed on your machine
- A JS package manager (pnpm recommended)
- Python 3 and pip (for backend services)
- Clone the repo:
git clone https://github.com/yourusername/linkedup.git cd linkedup - Install dependencies (using Turborepo + workspaces):
npm install
- Setup environment variables for each app (web, backend):
- E.g.,
.env.localfor Next.js,.envfor FastAPI, with your Supabase keys, Clerk keys, etc.
- E.g.,
-
Run the dev servers:
npx turbo run dev --parallel
This starts both the front-end (Next.js) and the FastAPI back-end.
-
Build:
npx turbo run build --parallel
-
Lint/test:
npx turbo run lint npx turbo run test
We employ a combination of xgBoost (for our “PTM” pipeline) and vector-based methods (via pgvector + OpenAI semantic embeddings) to produce high-quality matches. You can find our specific methodology in docs/MatchingPipeline.pdf:
- Integrate advanced ML algorithms for improved matching.
- Expand platform features (B2B meetings, mentor sessions, etc.).
- Enhance UI/UX for an even sleeker experience.
- Implement ML-based moderation for a safe community.
- Developed core real-time call functionality with Stream.
- Implemented secure and smooth onboarding with Clerk.
- Built a scalable backend on Supabase.
- Delivered a working demo showcasing our flagship features.
- API Integration: Mastering Clerk and Stream in a fast-paced hackathon environment.
- Team Collaboration: Effective communication and problem-solving under pressure.
- Scalability Concerns: Designing an app that's both scalable and modifiable for future growth.
- Real-Time Tech: Handling live calls and ML matching in a production-like setting.
- New Connection Modes: Explore B2B meeting opportunities, mentor-client sessions, and more.
- Enhanced Matching: Continuously refine our ML matching for even more effective networking.
- User Experience: Constant UI improvements to ensure a competitive, engaging platform.
- Community Building: Introduce gamified interactions and extended conversation features.