All work
AI infrastructure assistantPublic engineering build

Argus AI

A source-backed DevOps assistant with a working Kubernetes connector and an extensible tool layer for metrics, logs, and delivery systems.

Argus AIPublic build
01
EngineerAsks an infrastructure question in natural language.
02
Tool routerValidates intent and selects a read-oriented connector.
03
Trusted sourceReturns Kubernetes evidence before the model composes an answer.
NestJSTypeScriptKubernetes APIPrometheusLokiDockerRedis
Built forPlatform and DevOps teams
My roleCreator and infrastructure engineer
StatusPublic working project; Kubernetes connector implemented

Problem

Operational answers are slow when cluster state, metrics, logs, and deployment context live in separate tools.

What shipped

Natural-language requests routed through validated, read-oriented tools.

Working Kubernetes context for workloads, resources, and events.

Provider abstraction, validation, rate limits, safe logs, and local observability.

System shape

Argus AI

  1. 01

    Engineer

    Asks an infrastructure question in natural language.

  2. 02

    Tool router

    Validates intent and selects a read-oriented connector.

  3. 03

    Trusted source

    Returns Kubernetes evidence before the model composes an answer.

Evidence

Public source and setup pathWorking Kubernetes connectorDockerized observability stack

Tools & interfaces

NestJSTypeScriptKubernetes APIPrometheusLokiDockerRedis

Next honest step

Wire and verify Prometheus, Loki, and ArgoCD before presenting them as implemented integrations.

Discuss a similar system