WAEC-NECO AI Coach

Python + LLaMA.cpp + AI 2026

WAEC/NECO AI Coach 🎓

A lightweight, free AI learning assistant that helps Nigerian teens practice WAEC and NECO questions. Powered by the N-ATLaS GGUF model running locally via LLaMA.cpp, it generates clear, step-by-step explanations and supports multiple Nigerian languages... Yoruba, Igbo, and Hausa.

Designed specifically for Nigerian students seeking accessible, offline-friendly exam preparation.

Python LLaMA.cpp N-ATLaS GGUF AI Education Multilingual

Project Overview

WAEC/NECO AI Coach was built to tackle a real and critical problem: Nigerian students preparing for WAEC and NECO often have limited access to AI-assisted learning tools tailored to their curriculum.

The app generates practice questions using the N-ATLaS AI model and provides step-by-step explanations for each answer, making it easier for students to understand solutions, not just memorise them.

A key design decision was running the model fully locally using LLaMA.cpp and the N-ATLaS GGUF format. This keeps the app lightweight, fast, and offline-friendly, no cloud dependency or recurring API costs.

The interface is clean and minimal with two core screens - Onboarding and the AI Chat Coach - designed with teens in mind so there are no technical barriers to getting started.

Target Users

WAEC & NECO Students (Teens)

Timeline

Multi-phase Development

My Role

AI Backend Developer

Team

Solo Project

Deliverables

Local AI-Powered Learning App

Core Model

N-ATLaS GGUF (via LLaMA.cpp)

Languages Supported

Yoruba, Igbo, Hausa

License

MIT License

Development Process

Problem Identification

Recognized that Nigerian students preparing for WAEC and NECO lacked access to AI-assisted learning tools - with no lightweight, offline-friendly options that understand their curriculum or languages.

Model Selection & Local Deployment

Selected the N-ATLaS GGUF model for its Nigerian language support. Integrated it locally using LLaMA.cpp - a C++ inference engine - to eliminate cloud dependency and ensure the app runs fast and free.

App Architecture & UI Design

Structured the app around two focused screens: Onboarding (language/subject selection) and the AI Chat Coach. Used os and subprocess to manage local model execution and file handling cleanly.

Multilingual Support & Refinement

Extended the coach to respond in Yoruba, Igbo, and Hausa — making it accessible to students from different regions. Focused on a clean, teen-friendly UI that removes any friction between student and content.

Tools & Technologies

This project was built with a deliberately lean stack - prioritising local execution, speed, and zero cloud cost for end users.

Core Language

Python 3.10+

AI Inference

LLaMA.cpp N-ATLaS GGUF

System & File Management

os subprocess

Infrastructure

Local Model Hosting Offline-Friendly

Project Impact

A fully local, free AI coach that puts personalised exam prep in the hands of Nigerian students - no internet required, no cost barrier.

Nigerian Languages Supported

100%

Free Access

Cloud Dependency

∞

Scalability Potential

← Previous PriceCheck

All Projects

Next → Other Projects