Developed a Python-based FastAPI application to automate OT cybersecurity risk assessment. The API processes asset data (Excel) and vulnerability scans (PDF), enriching CVE information via the NVD API. It employs a hybrid approach: initial rule-based severity assignment followed by LLM refinement using Langchain with the Deepseek model via Groq. This provides nuanced risk levels and detailed justifications based on OT context (criticality, safety, isolation). The system outputs structured JSON results, streamlining vulnerability management in industrial environments.
Technologies: Python, FastAPI, Pandas, Pymupdf, Langchain, Groq API (Deepseek), NVD API.