AI Assistant Platform
RAG & Real-Time LLM Platform
Frontend Engineer · API Integration · Real-Time Streaming
- 2026
- Datamellon
- AI, Streaming
An AI assistant platform powered by RAG and LLM technologies. I worked on the client-side integration layer, connecting the frontend to streaming AI endpoints to create a real-time conversational experience.
My Contribution
I integrated RAG and LLM streaming APIs into the frontend, handling incoming streamed responses and presenting them progressively within the interface.
Instead of waiting for an entire response before displaying it, the interface could provide feedback as the model generated its response.
Technical Highlights
- Client-side API integration
- LLM streaming
- Server-Sent Events (SSE)
- RAG-powered responses
- Real-time UI updates
- Production debugging
- AWS environment troubleshooting
Production Challenge
During production deployment, LLM streaming encountered issues that did not appear in the development environment.
The problem involved differences between the local and production environments, including HTTP/HTTPS protocol requirements and infrastructure timeout behaviour.
How I Contributed
I worked alongside the Tech Lead and backend team to investigate the issue and update the client-side streaming implementation.
- Transition the client integration to secure HTTPS endpoints
- Update timeout handling
- Clean up legacy SSE proxy logic
- Validate the streaming behaviour in the production environment
What This Taught Me
The experience strengthened my understanding of real-time frontend systems and demonstrated how frontend behaviour can be affected by infrastructure, networking, and cloud deployment configuration.