• 

  AI finds best matches

• 

  Novus ai

Inspiration

Every Kazakhstani student knows the feeling — scrolling through scattered Telegram channels, random Instagram posts, and outdated university bulletin boards just to find one real opportunity. Most existing platforms focus on overseas programs, leaving local internships, jobs, and grants buried or ignored. I wanted one place that takes local opportunities as seriously as prestige global ones.

What it does

Damu aggregates internships, jobs, grants, scholarships, and competitions for Kazakhstani undergraduates — local and prestige global. Students fill in a quick profile (year, major, GPA, languages, goals), and Claude's AI scores every opportunity against their profile in real time, explaining exactly why they're a fit and where their gaps are. A Resources page also points students to platforms like Forage, Coursera, and Kaggle to help close those gaps.

How I built it

I built Damu using Bolt for rapid scaffolding, Claude's API for the AI matching logic, and deployed on Vercel. The data is currently a curated set of real opportunities we manually researched and verified, with a submission form for the community to contribute more.

Challenges I ran into

Environment variables not syncing properly between Bolt, GitHub, and Vercel cost us hours of debugging. I also hit a GPA normalization problem — Kazakhstani universities use different scales (4.0, 4.33, 5.0) — so we built a converter to compare students fairly. Caching stale AI match results when profiles changed was another tricky bug to track down.

Accomplishments that I'm proud of

Getting real AI reasoning working — not just a match score, but an honest explanation of fit and gaps — feels like the heart of the product. We're also proud of curating real, verified opportunities instead of placeholder data, and shipping a fully working, publicly deployed product end-to-end.

What I learned

Shipping is much harder than building. "It works on my machine" doesn't mean anything once deployment, environment variables, and caching get involved. We also learned that fairness matters in small details — like GPA scales — that are easy to overlook but matter a lot to real users.

What's next for Damu

I want to add an AI agent that scans local sources daily and flags new opportunities automatically, so opportunities find students instead of the other way around. I also want to connect a real database so community submissions go live after review, and grow Damu's opportunity list well beyond what one person can curate manually.

Built With

• bolt
• claude-api-(anthropic)
• github
• next.js
• react
• tailwind-css
• typescript
• vercel