AMResume
Back to featured work

Public prototype · developer tooling

Overwatch

AI-assisted incident triage tool for backend failures, root-cause analysis, blast-radius estimation, and postmortem generation.

TypeScriptNode.jsExpressDockerDeveloper ToolsIncident Triage
View GitHub Repo

Overview

Overwatch watches a local backend process, captures crash context, analyzes likely root cause, estimates blast radius, and generates a postmortem with verification and rollback guidance.

What I built

I built a local backend incident-triage prototype that observes process failure, recovers context, scores risk, and produces an incident artifact.

Technical focus

The technical focus is runtime-aware developer tooling: watchdog execution, crash parsing, workspace scanning, risk scoring, deterministic fallback analysis, and postmortem generation.

Why it matters

Overwatch shows how applied AI can support operational engineering without giving it unsafe write access by default.

System architecture

How the components connect

  1. 01

    Backend process

  2. 02

    Watchdog

  3. 03

    Crash parser

  4. 04

    Context & risk analysis

  5. 05

    Postmortem

Architecture is derived from the project's public documentation. It is a conceptual flow, not a live system diagram.

Engineering decision

Recover runtime context and suggest verification or rollback without allowing automatic production changes.

Failure mode

A retry or unreviewed patch can make an incident worse; a stack trace alone rarely captures the full context.

How to examine the evidence

Source includes a watchdog runner, workspace scanner, risk assessment and offline demo path.

npm run demo:offline
Source: watchdog runner

Prototype for local incident triage; not a proven production observability service.