Adaptive Behavioural Engine
APEX is a full-stack AI platform that uses retrieval-augmented generation (RAG) to deliver personalised behavioural coaching at enterprise scale. It combines vector search, role-based authentication, and a Kubernetes-deployed backend to serve thousands of concurrent users.
Large organisations struggle to deliver personalised behavioural coaching to thousands of employees. Traditional L&D programs are one-size-fits-all, expensive to scale, and lack measurable outcomes. There was no way to provide real-time, context-aware guidance that adapts to each employee's role, history, and growth trajectory.
Built a RAG pipeline with Qdrant vector search that ingests organisational knowledge and delivers contextual coaching. The system uses Keycloak for enterprise SSO, FastAPI for the backend, and React for the frontend, all orchestrated with Docker and deployed on Kubernetes.
- RAG pipeline with Qdrant vector search for contextual AI responses
- Keycloak integration for enterprise SSO and role-based access
- FastAPI backend with async processing for real-time coaching
- React frontend with responsive dashboards for admins and users
- Docker containerization with Kubernetes orchestration
- Scales to 10,000+ concurrent users with auto-scaling policies