Pragati

Inspiration

950+ scholarships exist in India. Most eligible students never apply — not because they don't qualify, but because the information never reaches them. Anjali from Meerut, Rahul from a government school, Vikram stuck in retail — they're not failing the system, the system is failing them.

What it does

PRAGATI is an AI-powered economic mobility platform that matches students to scholarships through conversational onboarding, then guides them step-by-step through eligibility, documents, and applications. A Claude-powered assistant answers questions in Hinglish, knows the user's profile, and makes intimidating government processes feel doable.

How we built it

FastAPI backend with PostgreSQL and a rule-based scholarship matching engine, paired with a React Native (Expo) frontend running as a web app. Claude API powers the context-aware AI assistant with per-user conversation memory stored in Redis. Deployed on Railway (backend) and Vercel (frontend).

Challenges we ran into

There is no centralized scholarship database in India — we had to manually curate and structure data from NSP, buddy4study, and state government portals into a queryable schema. Getting Claude to respond accurately in Hinglish while staying grounded in the user's specific profile (and not hallucinating scholarship details) required careful system prompt engineering.

Accomplishments that we're proud of

The onboarding flow — a 12-step conversation that feels nothing like a form — extracts a complete user profile and immediately returns a ranked, personalized scholarship list with match reasons. That full loop, from zero to "here are 8 scholarships you're eligible for and why," working end to end, is something we're genuinely proud of.

What we learned

Economic mobility problems are fundamentally information problems. The scholarships exist, the money exists, the eligibility exists — what's missing is the last-mile connection. Building that connection is less about algorithms and more about trust, language, and meeting people where they are.

What's next for First Hack

Document OCR to auto-fill scholarship application forms from a photo. A WhatsApp bot for users without smartphones. A verified outcomes database where past students confirm placements, making course recommendations trustworthy. And expanding the scholarship database from 25 to 950+ with automated daily scraping.