ScholarshipHub

How I Got Frustrated With Scholarships

So I'm sitting with my friends one day, and one of them mentions this scholarship. Fully funded. $100K. To study abroad. And I'm like... wait, how did you find that?

He shows me. It's just... scattered. Different websites. Different languages. Some are in Arabic, some are in English, some are on random forums. He spent like 4 hours finding this ONE scholarship.

And then I realized: he's lucky he found it at all.

Most of my friends? They have no idea these even exist. And the ones who try to apply? They write applications that get rejected. Not because they're not smart. But because they don't know what to write or how to make it strong.

That's when I got frustrated. Like... why is accessing help this hard?

What it does

Three things:

Finds your scholarship matches - Input your profile, get 5 scholarships that actually fit you (not 500 random ones)
Reviews your documents - Upload a CV. AI tells you what's working and what needs fixing. Actionable feedback, not generic stuff.
Tracks your progress - See your CV score improve from Draft 1 to Draft 2. When you see improvement, you keep going.

Plus: A roadmap for each scholarship so you don't miss anything.

How we built it

Solo. Next.js + Claude AI + Supabase.

Started by surveying 44 students (what's actually hard?). Built based on their answers, not my assumptions.

Tested with real people. They said "This doesn't work." I fixed it. Did it again.

Didn't polish for weeks. Got it working fast and moved on.

Challenges we ran into

AI feedback was generic at first. Spent hours rewriting prompts until Claude gave actually useful feedback.

Ranking 2,000 scholarships by fit was hard. Age, country, education level, funding type... Get one weight wrong and everything breaks. Tested with real people. Adjusted. Tested again.

Time. I'm a Grade 10 student with school and other stuff. Had to cut every nice-to-have and focus only on what matters.

Accomplishments that we're proud of

Built a complete app that actually works.

Got 44 real students to validate it (91% said they'd use it).

Made the AI feedback actually useful, not generic.

Proved I could build something from idea to live product solo.

What we learned

Validation first. Talk to people before you build. It saves time and prevents building the wrong thing.

Focus beats perfection. Moving fast is better than polishing for weeks.

AI is a tool. It only works if you engineer it right. And humans should always make the final decisions.

The real problem isn't technology. It's access. A kid in rural Egypt shouldn't have fewer opportunities than a kid in Cairo just because they don't know the system.