DESCRIPTION
- Final-semester Research Project thesis that received the Best Research Paper Award in QUT.
- Focused on optimizing RAG for cybersecurity. When handling incidents, internal and up-to-date information is critical; standalone AI models (e.g., ChatGPT, Gemini) cannot access private or latest sources, which limits their effectiveness.
- Investigated whether RAG can effectively answer the latest CVEs—high-priority information for cybersecurity response.
- The system built in this research performed better on recent CVE questions compared to using GPT-4o alone.
TECH STACK
Language
Python
Others
FastAPI, text-embedding-3-small, Chroma DB