Overview
The DJKYD team (David Da Graca, Jacob Abraham, Kamji Shehu, Yasmeen Jabi, Dominic Murphy) developed an AI Support Hub as our final-year capstone project in the Bachelor of Information Technology – Information Resource Management stream. This web-based platform offers Carleton University students and faculty centralized guidance on academic, ethical, and practical use of artificial intelligence. The aim was to foster informed, responsible engagement with AI technologies in learning and teaching environments.

Problem Statement
With AI tools rapidly entering classrooms, both students and instructors are left unsure about what’s permitted, what’s ethical, and how to use these tools responsibly. The existing Carleton documentation was either scattered, outdated, or overly technical. Our project addressed this gap by creating a user-centered website focused on information clarity, digital literacy, and inclusive design.

Key Components
1. AI Resource Hub Website
- Built using HTML, CSS, JavaScript, PHP, Flask (for chatbot integration), and MySQL.
- Provides simplified guides on AI ethics, plagiarism, citation tools, and assignment-specific AI considerations.
- Includes dynamic search, category filters, and accessibility enhancements to align with WCAG standards.
- Designed responsively to support both desktop and mobile users.

2. Community Forum (phpBB)
- Fully configured and branded phpBB forum hosted on our server.
- Organized into categories (e.g., Coursework Help, AI Tools Q&A, Policy Discussions).
- Integrated with MySQL for user registration, moderation, and post management.
- Seeded with over 20 sample posts to simulate activity and show functionality during demo sessions.
- Accessible only to Carleton students via a custom login script and email verification.

3. AI Chatbot Prototype
- Developed using a lightweight Flask application to handle user input and generate responses locally.
- Backend written in Python, connected to a small language model hosted via LM Studio on the same machine.
- Frontend built with basic HTML, CSS, and JavaScript, enabling real-time interaction through a chat window embedded in the homepage.
- Uses a simplified model to ensure reasonable response times without needing advanced GPU hardware.
- Current setup includes preloaded FAQ data; future development will integrate Carleton-specific policy documents using vector-based semantic search.


4. Server Deployment & Security
- Hosted on a DigitalOcean VPS running Ubuntu 22.04.
- Installed and configured full LAMP stack: Linux, Apache, MySQL, and PHP.
- All source files managed via GitHub with version control and CI/CD using GitHub Actions.
- HTTPS enforced using Certbot and Let’s Encrypt SSL.
- phpMyAdmin used for database management and monitoring.

5. Backup & Recovery Plan
- Automated server backups implemented using `rsync` and `cron` to a secure secondary volume.
- Separate daily database dump via shell script into a compressed archive folder.
- Manual backup checkpoints maintained after each major feature push for recovery testing.
- Disaster recovery simulated during testing week—successfully restored to a previous build after database corruption.

Challenges & Lessons Learned
- Initial chatbot integration was more complex than expected due to token limits and API restrictions; we pivoted to a local model for demo purposes.
- Styling phpBB to match our website took trial-and-error with template overrides and CSS tweaks.
- Server setup exposed us to real-world troubleshooting (e.g., port conflicts, database timeouts).
- Our biggest takeaway: always plan double the time for deployment and testing—final week bugs reminded us of the importance of buffer time.

Future Outlook
- Train chatbot on actual Carleton policies and connect to course-specific prompts.
- Expand forum moderation tools and introduce SSO using Carleton credentials.
- Build faculty-only sections for discussions on academic policy and grading with AI.
- Conduct broader UX testing with real Carleton users and partner with the university’s TLSS or Library for sustainable adoption.


Conclusion
This capstone was more than a technical build—it was a community-centered design initiative. We created a solution that bridges knowledge gaps, promotes responsible technology use, and can grow into something with long-term institutional value. Our work reflects not just technical skill, but also our passion for clarity, inclusivity, and digital leadership.