1 / 19
C-Link - Construction SaaS
Problem
Client onboarding for a construction SaaS platform depended on manually turning contractor and quantity-surveyor documents into dynamic templates, slowing delivery for clients including McLaren.
Solution
Led engineering delivery end-to-end and, as a proof of concept directly assigned by the CTO, architected a multi-agent AI pipeline that automates the document-to-template workflow - cutting manual processing time by 40%. The platform itself is built on a custom PHP framework (Slim, with fat services) in a microservices architecture, with React and Angular frontends.
What I built
Platform Infrastructure & Integrations
Documents, async workflows, event notifications, and credentials needed a secure, event-driven backbone, alongside integration with Asite for construction document management.
Used AWS S3 for document storage, SQS for queuing and async behaviour, SNS for event-based email notifications, and Secrets Manager for credentials management. Integrated Asite for construction document management and implemented a third-party socket-powered real-time notifications system.
Engineering Practices & DevOps
The team needed consistent containerisation, testing, code quality, and deployment practices across services.
Active use of Docker and GitHub Actions CI/CD, PHPUnit test coverage, and SonarQube code quality gates, running on AWS ECS and EKS (EKS managed by the DevOps team).
Framework Error Handling Standardisation
The custom PHP framework lacked a consistent exception-handling system in core API communication - a debugging headache that had persisted for 7 years.
Standardised error handling across the custom framework by implementing a proper exception system for core API communication and adding structured logging, resolving the long-standing debugging problem.