AI Systems Engineer · RAG & Agentic AI · Full-Stack ML
Final-year Computer Engineering student at the University of Ruhuna, building AI systems that run in production — not just notebooks. Focused on Retrieval-Augmented Generation, multi-modal knowledge graphs, and scalable ML pipelines.
- Currently building agentic RAG systems and multi-modal knowledge graphs (GraphRAG)
- Researching Graph Neural Networks + blockchain for vehicular networks (SDVNs)
- Open to collaborate on AI/ML research, computer vision, and open source
| Project | Description | Stack | Status |
|---|---|---|---|
| NeuroSpace | Multi-modal GraphRAG agent that turns video lectures and papers into a navigable knowledge graph with agentic retrieval. | LlamaIndex · Neo4j · FastAPI · Docker | in development |
| ContextIQ · ↗ | Chat with your PDFs via a Gemini + Pinecone RAG pipeline — upload, chunk, embed, retrieve, answer. | Python · Streamlit · Gemini · Pinecone · LangChain | completed |
| AI Data Analyst · ↗ | Natural-language CSV analysis powered by Groq + LangChain agents. | Python · Streamlit · LangChain · Groq · Pandas | completed |
| Salary Estimator · ↗ | End-to-end MLOps pipeline with CI/CD, containerization, automated testing, and deployment. | Python · Flask · Docker · GitHub Actions | completed |
| EV Charging Booking · ↗ | Real-time EV charging booking with route optimization and live station monitoring. | React Native · Node.js · Socket.io | completed |
| Live Talk · ↗ | End-to-end encrypted real-time chat with secure file transfer and multi-user channels. | React · Redis · Socket.io | completed |
| Enterprise MCP Agent | Secure MCP server bridging agentic LLMs with relational POS databases via native tool-calling and domain guardrails. | Python · MCP SDK · SQLite · Groq | in development |
| Voxis | Voice assistant that records, transcribes with Whisper, and responds conversationally with Llama. | Python · Groq · Whisper · NumPy | in development |
Graph Neural Networks · Knowledge Graph Construction · Multi-Agent AI Systems · Blockchain for Vehicular Networks
