SkillLink

• 

  adding profile

• 

  searching for the required skill to learn

• 

  minimal landing page

SkillLink is a lightweight platform that helps students find the right people to learn from or collaborate with. Every campus has a huge amount of hidden talent, but most students never meet the people who share their skills or interests. SkillLink makes this process simple by connecting students based on what they know and what they want to learn.

Users enter their name, batch, skills they are confident in, and skills they want to pick up. The system cleans and standardizes these skills using AI, then stores everything in a MongoDB Atlas database. When someone searches for a skill, the platform doesn’t rely on exact keywords. Instead, it uses vector embeddings to understand related concepts. For example, searching “frontend” can match people who know React or Vue. The platform also generates two match lists for every user: people they can teach, and people who can teach them.

The backend is built with Flask, handling all API routes and matching logic. MongoDB Atlas stores user profiles along with their vector representations. The AI components handle skill extraction and embedding generation. The frontend is a simple, responsive interface built with HTML, CSS, and JavaScript.

One of the interesting parts of the project was implementing cosine similarity to compare skill vectors. This allowed the system to understand relationships between skills without relying on a predefined dictionary. We also had to solve challenges like normalizing different ways students describe the same skill and making the local server accessible across devices during testing.

We’re proud that we were able to build a complete AI pipeline — from extraction to embedding to semantic search — within the event timeframe. We also learned a lot about vector databases, prompt design, and building a clean development setup for live demos.

In the future, we’d like to move to MongoDB’s native vector search for better performance, add real‑time chat so matched users can connect instantly, and introduce a skill verification system where students can endorse each other.

Built With

• and
• css3
• flask
• html5
• javascript
• mongodb-atlas
• numpy
• openai-api-(gpt-4o-mini-&-embeddings)
• python