Inspiration
I saw how many students in Pakistan struggle with English-based learning apps. They understand concepts better in Urdu, but there are no good AI tutors for Urdu. I wanted to make education accessible in our native language using Meta Llama.
What it does
Urdu AI Tutor helps students learn math, science, and other subjects by asking questions in Urdu. The app uses Meta Llama 3.1 to give instant answers, step-by-step explanations, and practice problems - all in Urdu.
How we built it
- AI Model: Meta Llama 3.1 for Urdu language understanding and Q&A
- Frontend: React for the student interface
- Backend: Python + FastAPI to connect to Llama
- Prompts: Carefully designed to make Llama respond accurately in Urdu
Challenges we ran into
- Making Llama give consistent and accurate answers in Urdu
- Designing prompts so the AI explains concepts simply for students
- Testing with real Urdu questions to improve responses
Accomplishments that we're proud of
- Built a working AI tutor that speaks fluent Urdu
- Successfully integrated Meta Llama 3.1 for education
- Created an interface that any Urdu-speaking student can use easily
What we learned
I learned how to work with LLMs, design good AI prompts, and build for real user problems. Most importantly, I learned that language should not be a barrier to education.
Built With
- ai
- edtech
- education
- llama
- meta
- natural-language-processing
- python
- urdu


Log in or sign up for Devpost to join the conversation.