Rifqa: offline English tutoring AI for Algerian students with Arabic explanations (ADTC 2026 submission)
# ADTC 2026 — Submission Template
This is the official template repository for the **Africa Deep Tech Challenge 2026** Laptop LLM track.
Fork this repository, fill in the required files, and submit your repository URL via adtc-2026.devpost.com.
---
## ✅ Submission Checklist
Before submitting, confirm every item:
- [ ] Your repository is **public** on GitHub
- [ ] `metadata.json` is fully filled in — no placeholder values remain
- [ ] `metadata.json` contains exactly **2 test prompts** in the `test_prompts` array, written for your chosen domain
- [ ] `download_model.sh` successfully downloads your model to `model/`
- [ ] The downloaded file is a valid **GGUF format** (`.gguf`) weight file
- [ ] `model/*.gguf` is listed in `.gitignore` — do **not** commit large weight files
- [ ] `REPORT.md` is filled in with your technical writeup
- [ ] Running `bash download_model.sh` completes without errors
- [ ] Your model runs entirely **offline** — zero external network calls during inference
---
## 📁 Required File Structure
```
your-submission/
├── metadata.json ← Required. Team, model, and test prompt metadata.
├── download_model.sh ← Required. Downloads your .gguf model weight file.
├── REPORT.md ← Required. Technical writeup (problem, design, benchmarks).
├── model/
│ └── your-model.gguf ← Downloaded by the script above. Do NOT commit.
└── .gitignore ← Must exclude *.gguf and model/ from version control.
```
---
## 📝 metadata.json
Fill in every field. No field should remain at its placeholder value.
```json
{
"team_id": "your-team-id",
"domain": "coding_assistants",
"language_scope": ["en"],
"african_alpha_claim": false,
"budget_laptop_claim": true,
"submitter": {
"name": "your-name",
"email": "your-email@domain.com",
"github_handle": "your-github"
},
"cross_disciplinary_pairing": {
"discipline": "education",
"load_bearing": true,
"description": "Brief description of how your model serves a real-world domain."
},
"test …