How a freelance CMS project evolved into an AI powered platform on AWS. A look at the enterprise level features from a ticketed support system to an LLM co pilot for admins and why this £15-20k project was about more than just money.
Following my posts about navigating changes in my day to day role, I decided the best way to stay energised and keep my full stack skills razor sharp was to dive back into freelance work. The perfect opportunity appeared, a chance to build a complete, ground up Content Management System (CMS) for a group that organises a fantastic series of race events.
Leveraging a modern tech stack of React, TypeScript, and a cloud-based AWS backend defined with Infrastructure as Code (IaC), this project quickly evolved into more than just a CMS it became an intelligent, AI-powered platform.
Building an Enterprise Grade Solution
The goal was to give the administrators total control and to make the user experience as seamless as possible. I went all out to build something truly robust and future-proof.
I'm incredibly proud of the feature set we packed into this platform:
Dynamic Event Management: Admins have full control to create new events, series, and individual races, and can upload custom images for each.
Advanced Team System: The system allows for team creation where designated "Captains" can upload their own team logos, which then enter a moderation queue for admin approval.
A Powerful Admin Dashboard: The heart of the system is a full admin dashboard with various role-based access levels, ensuring team members only see what's relevant to their role.
Flexible Data Ingestion & Enrichment: I built two distinct ingestion methods: pasting CSV data for historical results and using webhooks for future updates. The system also enriches this data by connecting to external APIs for supplementary information.
Integrated Support System: To streamline communication, I built a messaging system that functions like a support desk, allowing users to create tickets to talk directly with admins and moderators.
AI-Powered Assistance (LLM Integration): This was where I really pushed the boundaries. I integrated a Large Language Model to act as a dual-purpose assistant:
For Users: A help bot that not only answers questions but is trained to detect signs of frustration in user queries. If it senses a user is struggling, it can proactively offer advice or even create a support ticket automatically.
For Admins: An "Admin Co pilot" to assist them in two key ways. Firstly, it acts as an interactive guide, providing step by step instructions on how to use various admin systems, such as the score importing or historical data import functions. Secondly, it can perform complex tasks for them. For example, they can ask it to work out a complicated scoring formula in natural language, like "calculate points for the top 50 finishers where first place gets 100 points, and the rest get points in batches of 5 places down to 50th," and the system will generate the correct data.
More Than Just a Project
A project of this scale, especially with the custom AI integration, would typically be valued in the £15,000 - £20,000 range. However, this project was different.
Firstly, it was a personal thank you to the company owner, who has helped me out in the past. But just as importantly, it was a strategic investment in my own craft. It served as a powerful demonstration of my ability to architect and deliver a complete, complex application from scratch. I went above and beyond the requirements because I was passionate about the problem and wanted to build a portfolio piece that truly represents the quality of work I strive for.
Next Steps
The system is now built, tested, and ready. The finish line is in sight! The site isn't ready to go live just yet, as the organisers are currently in the process of adding all their historical data. I'm incredibly excited to see it launched and to watch it become an essential tool for their community.