Sangamesh.Dandin
All projects

Case 06 / 09Threat intelligenceCompleted

ZEROTRACE AI.

Threat intelligence with exploit prediction.

A threat-intelligence platform that merges CVE feeds from NVD, CISA KEV and ExploitDB into one risk-ranked view. An XGBoost model estimates how likely each vulnerability is to be exploited, the Groq API summarises advisories, and FAISS powers semantic search. Streamlit, FastAPI, PostgreSQL and Docker.

What I built

  • Built ingestion for CVE feeds from NVD, CISA KEV and ExploitDB into one risk-ranked pipeline.
  • Trained the XGBoost exploit-prediction model on CVSS score, exploit availability, CWE category and vendor data.
  • Added FAISS semantic CVE search and Groq-generated advisory summaries.
Problem
CVE overload
Built with
XGBoost, FAISS
Result
Exploit risk ranking
Links
  • XGBoost
  • FAISS
  • CVE
  • Streamlit
  • PostgreSQL
  • Docker

01Overview

At a glance.

06 / 09 · THREAT INTELLIGENCE ZeroTrace AI 01Ingest02Predict03Summarise04Search PIPELINE · 4 STAGES

02Under the hood

How it works.

  1. 01

    Ingest

    CVEs from NVD, CISA KEV and ExploitDB

  2. 02

    Predict

    XGBoost estimates exploit probability

  3. 03

    Summarise

    Groq API summarises advisories

  4. 04

    Search

    FAISS vector search across the corpus