
Club member portal
Club Member Portal is a web-based platform for AWS Student Builder Group members to sign up, log in, reset their passwords, and interact with an AI chatbot. The chatbot answers member questions using the provided club documents and displays the source of each answer. If the required information is not available, it provides a safe fallback and directs the member to the campus AWS Student Builder contact.
Introduction
Our team developed a Club Member Portal for AWS Student Builder Groups.
The application allows members to create accounts, log in, reset passwords, and use a chatbot to ask questions about the Student Builder Group.
The core AI requirement is that answers must be based on the provided club documents, with the relevant source shown for each response.
My Role — Backend and AI Development
I worked primarily on the backend and AI functionality of the project.
My responsibility was to connect the user-facing application with authentication, document processing, retrieval, chatbot functionality, and safe response handling.
My Contributions
- Developed the backend/API functionality
- Implemented authentication-related backend logic
- Implemented password-reset flow
- Loaded the 8 starter documents
- Implemented document retrieval
- Connected the chatbot to the retrieved document content
- Implemented source information in responses
- Implemented fallback handling
- Connected the frontend with backend APIs
- Worked on the AWS deployment architecture
Backend Architecture
The basic application flow is:
Frontend
↓
Backend API
↓
Authentication / User Management
↓
Document Retrieval
↓
AI Chat Processing
↓
Response + Source
The backend acts as the layer connecting the frontend with authentication and AI functionality.
Authentication
The portal supports:
- Email/password registration
- Login
- Forgot-password functionality
- Password reset
The challenge specifically requires email/password authentication and a password-reset flow.
For a local prototype, the challenge allows tools such as Nodemailer for sending password-reset emails.
Starter Document Pack
The chatbot uses the eight documents provided for the challenge:
- "01-onboarding-faq.md"
- "02-aws-account-setup.md"
- "03-builder-center-publish.md"
- "04-bedrock-starter.md"
- "05-hackathon-rules.md"
- "06-workshop-index.md"
- "07-lambda-patterns.md"
- "08-sbg-community.md"
The challenge requires the starter pack to be loaded when the application starts.
RAG-Based Question Answering
The chatbot follows a retrieval-based approach.
The flow is:
User Question
↓
Search the Club Documents
↓
Retrieve Relevant Content
↓
Provide Relevant Context to the AI
↓
Generate Answer
↓
Return Answer + Source
This keeps the chatbot focused on the information contained in the club document pack.
Source Citations
Every chatbot answer should include its source.
For example:
Answer:
Information retrieved from the club documentation.
Source:
"06-workshop-index.md"
Section:
Relevant workshop section
The challenge specifically requires the chatbot to cite the source file and section.
Safe AI Fallback
A major part of the backend logic is handling questions that cannot be answered from the documents.
If relevant information cannot be found, the system should not guess.
Instead, it returns a fallback response telling the member that the information could not be found and directing them to the campus AWS Student Builder contact.
This follows the challenge's requirement to avoid hallucinated answers and not invent AWS pricing, limits, or policies.
Data and Security
The application should avoid exposing sensitive information.
Passwords should not be logged or stored as plain text, and private student information should not be unnecessarily logged. This follows the security guidance in the challenge document.
AWS Deployment Architecture
For the local demo, the application can run locally because live AWS deployment is optional.
For a production AWS architecture, the planned services are:
- Amazon Cognito — Authentication
- Amazon SES — Password-reset emails
- Amazon S3 — Document storage
- Backend/Chat API — Application logic
- Amazon Bedrock — AI responses
The challenge specifically recommends explaining these components as part of the AWS deployment plan.
What I Learned
Working on the backend and AI side helped me understand:
- Backend API development
- Authentication flows
- Password-reset handling
- Document-based question answering
- RAG concepts
- Source citation
- AI safety and fallback handling
- Integrating AI functionality with a web application
- Mapping application components to AWS services
Future Improvements
Future improvements could include:
- Per-user chat history
- User feedback on AI answers
- Improved document retrieval
- Complete live AWS deployment
- Better monitoring and analytics
Chat history and thumbs-up/down feedback are listed as optional features, while live AWS deployment is a bonus.
Conclusion
The backend and AI layer provides the functionality behind the Club Member Portal.
By combining authentication, document retrieval, AI-generated responses, source citations, and safe fallback handling, the project provides a focused AI assistant for AWS Student Builder Group members.
The project also demonstrates how the same application can move from a local hackathon prototype toward an AWS-based architecture using services such as Cognito, SES, S3, and Amazon Bedrock.
Build. Learn. Collaborate.
Enjoyed reading this content? Let the author know!
Your likes, comments, shares, and saves help creators reach more builders.
Loading recommendations
Loading article