AI Log Analyzer

How it started

Log analysis is one of those tasks that's simultaneously critical and mind-numbing. A NOC operator staring at thousands of log lines per hour will miss things — not because they're bad at their job, but because humans aren't built for that kind of pattern recognition at volume.

I wanted to build something that could do the first pass: ingest raw log data, identify what's normal, flag what's anomalous, and present it in plain language. Not a replacement for a trained analyst — a force multiplier for one.

How it works

• Analysis — a Python script runs Claude against raw syslog or Windows Event Log files and outputs structured JSON
• Dashboard — the frontend reads that JSON and renders a full threat assessment: severity breakdown, attack chain, anomaly detection, recommended actions
• Simulation mode — a pre-generated cyber attack scenario plays out in real time, showing how the dashboard responds as an attack progresses

Built as a portfolio demonstration of AI-assisted log triage. The analysis pipeline is functional — the frontend is the deliverable here.

Tech stack

Status

Live. The demo runs on pre-analyzed log datasets including a real-time cyber attack simulation. The backend Python analyzer runs locally against real log files — the frontend is fully deployed on Netlify.

Want a custom version for your team? Get in touch.