EN · FR
Languages
Llama 3.3 70B
Model
Deployed
Status
Overview
A bilingual (French/English) open-source AI chatbot providing career guidance, government programme information, emotional support, and local resource recommendations to young Cameroonians, auto-detecting language and responding naturally in either.
The Challenge
Cameroonian youth faced a real gap in accessible, culturally-relevant career and emotional-support guidance, particularly in French, with most AI tools reliably serving only English speakers.
The Solution
Built on Django 5 and DRF with a dedicated FastAPI AI microservice calling Groq's Llama 3.3 70B, plus Tavily for live search on government programmes and local resources. Implemented crisis detection via pre-LLM keyword scanning so distress signals are caught before they ever reach the language model, paired with location-based resource recommendations.
Outcome
Fully deployed on Dokploy under a custom domain with a production-tested backend. Bilingual auto-detection means users get coherent, contextually relevant responses without ever switching modes manually. Open-source so local organisations can fork and extend it for their own communities.
Stack
Have a similar project in mind?