2026 / Full-stack / AI systems

Rootify

An AI-driven incident analysis tool that ingests logs from 30+ microservices, correlates transactions, and generates root-cause analysis using LLMs.

JavaSpring BootPythonReactPostgreSQLElasticsearchLLMsVector DB

Context

Rootify is a portfolio case-study project for incident triage across distributed services.

Problem

Production incidents often require engineers to jump across logs, dashboards, service ownership, and partial business context before forming a hypothesis.

Approach

  1. Ingest service logs and correlate them through transaction IDs and known request boundaries.
  2. Use Elasticsearch for precise event retrieval and a vector store for semantically related historical context.
  3. Generate role-aware explanations so engineers get technical traces while business users get customer-impact summaries.

Impact

  • Demonstrates a practical RAG architecture beyond a document-chat demo.
  • Shows how LLMs can support debugging while keeping source evidence visible.
  • Creates a strong portfolio story around AI agents, observability, and backend systems.

Sample incident flow

A checkout transaction starts failing after a downstream timeout. Rootify groups logs, identifies the failing dependency, retrieves similar incidents, and drafts the first RCA.

01 Upload or stream logs02 Correlate transaction path03 Retrieve similar incidents04 Generate RCA draft
Back to Projects