---
title: "AWS AI SOC Analyzer — Paolo Ronco"
language: "en"
url: "https://paoloronco.it/en/projects/aws-ai-soc-analyzer"
---

[← projects](/en/projects)

# AWS AI SOC Analyzer

Experimental AI-assisted SOC triage project integrating n8n, AWS API Gateway, Lambda, and Bedrock to analyze security alerts and return structured results.

Security

Problem

I wanted to learn AWS through a use case close to my work: adding context to SOC alerts without handing security decisions to an LLM.

Solution

I connected n8n to an IAM-authenticated API Gateway route; Lambda normalizes the alert, calls Nova 2 Lite on Bedrock and validates the JSON response.

Outcome

V1 verified from n8n through AWS and back with a sample SSH alert; connecting the live Splunk webhook is the next step.

[GitHub repository ↗](https://github.com/paoloronco/aws-ai-soc-analyzer)[Read the article ↗](https://paoloronco.it/en/writing/aws-ai-soc-analyzer-imparare-aws-con-un-progetto-soc/)

## Architecture and boundaries

In the intended flow, Splunk produces the alert, n8n sends it in a SigV4-signed request, API Gateway exposes `POST /analyze`, and Lambda calls Bedrock using its own IAM role. n8n receives severity, confidence, evidence and suggested actions in validated JSON.

V1 uses a manually triggered n8n workflow and an SSH-oriented input schema. The model’s analysis remains an aid to the analyst; operational decisions require deterministic rules and human review.
