• 

  FaithLearnAI Logo

Inspiration: FaithLearn AI was inspired by a simple question:

What if every learner could have an intelligent learning companion that understands where they are and knows what they should learn next?

Through my work in education and community development, I have seen that many learners have access to educational content but do not always have access to personalized guidance, continuous feedback, or a personal tutor.

I wanted to build something beyond a traditional chatbot — an AI learning companion that could understand learner context, teach, create practice activities, evaluate performance, track progress, and guide the learner toward their next step.

That idea became FaithLearn AI.

What it does: FaithLearn AI is an agentic AI learning companion that turns a simple conversation into a personalized learning journey.

A learner can use FaithLearn AI to:

• Ask for personalized lessons.
• Get explanations appropriate to their learning level.
• Generate quizzes and practice activities.
• Submit answers for grading and feedback.
• Check their learning progress.
• Receive recommendations for what to learn next.

The key idea is that FaithLearn AI does not stop at providing an answer.

It creates a learning loop: Understand → Teach → Practice → Evaluate → Track → Recommend.

This makes FaithLearn AI more than a conversational AI tutor. It is designed to support the learner throughout the learning process.

How we built it:

FaithLearn AI was built as a full-stack AI web application.

Frontend

• React
• Vite
• Responsive web interface
• Learner profile and personalized experience

Backend

• Node.js
• Express.js
• REST API

AI

• Amazon Bedrock
• OpenAI model through Bedrock for AI reasoning and tutoring

Agentic tools

We implemented Model Context Protocol (MCP) to connect the AI reasoning layer with educational tools, including:

• get_student_progress
• get_lesson
• create_quiz
• grade_answer

These tools allow the AI to move beyond generating text and perform meaningful learning actions.

Authentication and deployment

• Supabase Auth for learner authentication
• Google, Facebook, and email authentication
• Render for cloud deployment The result is a live, accessible AI learning experience rather than only a local prototype.

Challenges I ran into: One of my biggest challenges was connecting multiple systems into one reliable experience.

I had to make the React frontend, Express backend, AI service, MCP tools, authentication, environment variables, and cloud deployment work together correctly. Authentication was another major challenge. During development, the application was automatically signing users out shortly after login. I had to carefully troubleshoot the authentication state and frontend logic until the login session remained stable. I also encountered differences between the local development environment and the deployed cloud environment. This required testing API URLs, environment variables, authentication configuration, and production behavior. These challenges helped us understand that building an AI product is not only about the AI model — it is about making the entire system reliable for the user.

Accomplishments that I am proud of: I am proud that FaithLearn AI evolved from an idea into a working, deployed AI learning application.

Some of my key accomplishments include:

• Built a functional AI tutoring experience.
• Implemented personalized learner context.
• Integrated AI reasoning with educational MCP tools.
• Created lesson, quiz, grading, and progress capabilities.
• Implemented learner authentication.
• Successfully deployed the application to the cloud.
• Created an end-to-end learning workflow.
• Designed the system around an agentic approach rather than a simple chatbot. Most importantly, I demonstrated that an AI assistant can move from: Answering a question → taking a learning action → understanding progress → recommending the next step.

What we learned: Building FaithLearn AI taught me that effective AI applications require more than a powerful language model.

I learned how to connect AI reasoning with real application tools and user context.

MCP was particularly important because it provided a structured way for the AI to interact with educational capabilities instead of simply generating responses. I also learned valuable lessons about:

• AI application architecture
• MCP tool design
• React and Node.js integration
• API development
• Authentication
• Cloud deployment
• Environment configuration
• Production debugging
• Designing AI around real user needs Most importantly, I learned to think about AI as an agent that can help accomplish a goal, rather than simply as a system that generates text.

What's next for FaithLearn AI: FaithLearn AI is only the beginning.

My next steps include:

• Expanding to more subjects such as Mathematics, English, and Science.
• Improving curriculum alignment.
• Developing richer learner analytics.
• Improving personalized recommendations.
• Expanding learning activities and assessment tools.
• Supporting more learners and educational communities.
• Exploring mobile and voice-based learning experiences.
• Building stronger partnerships with schools and community education programs.

Our long-term vision is to make personalized learning support more accessible, especially to learners who may not have access to a private tutor. FaithLearn AI — Learning that understands context, empowers every learner, and guides what comes next.

Built With

• css
• express.js
• github
• html
• javascript
• mcp
• node.js
• openai
• react
• rest-api
• supabase
• vite