Edumatch AI

• 

  Opportunities presented to a user by our AI based on their gpa, skills, needs, domain, etc.

EDUMATCH__AI was inspired by the frustrating gap in educational opportunities I faced as a computer science student, where platforms like HackerEarth list scholarships, hackathons, and internships but lack smart personalization. Drawing from my own struggles with overwhelming options and poor matching, I envisioned an AI-powered platform that uses explainable algorithms to connect student skills with the best opportunities, much like recommendation systems in Netflix but tailored for education .

Through this hackathon project, I learned to build full-stack AI applications from scratch, mastering vector embeddings for semantic matching and optimizing recommendation engines under tight deadlines. I gained deep insights into user-centric design, integrating real-time features like dynamic filtering by skills, location, and eligibility, while balancing AI accuracy with low-latency performance using hybrid search techniques .

Challenges included repository access issues from private settings, resolved by local cloning, and rate-limiting from 30+ opportunity APIs, fixed with async batching and exponential backoff. Balancing precise AI matching with fast responses required caching and hybrid keyword-semantic search .

Built With

• javascript
• json
• react
• rest
• typescript