ARKONA

Jhon Arango

Agentic AI Systems Engineer

Every organization is rushing to deploy AI.
Nearly 80% of AI projects fail.

— RAND Corporation, 2024: "More than twice the failure rate of IT projects that don't involve AI"

The reason? Organizations buy AI tools without a methodology for integration. No structured analysis of which tasks AI should own. No accountability framework. No delegation governance. They skip the hardest question: who is responsible when the AI makes the wrong call at 3 AM?

I built the system that solves this.

COMET identifies where AI belongs

Decomposes any domain — aircraft maintenance, healthcare, finance — into organizational structures, job roles, and tasks. Then classifies each task across 5 delegation levels from fully human to fully autonomous, grounded in 20 industry standards. The output: an auditable RACI matrix with the AI agent as a first-class participant.

ARKONA builds and operates the agents

A production software factory that builds, deploys, and monitors AI agents with the same rigor as mission-critical defense systems. Rigorous logging, circuit breakers, provenance signing, and human oversight at every level. Not a demo — a system that runs at 3 AM and can be explained to regulators.

Services

AI Agents

928

Commits

View My Work Contact Me

What I Build

Multi-Agent Orchestration

Purpose-built agent harnesses where specialized AI agents collaborate on complex tasks — with structured communication, shared memory, and human-in-the-loop governance at every stage.

Hybrid LLM Routing

Intelligent model selection that dynamically routes between cloud and local models based on task complexity, context requirements, and cost constraints — optimizing for both capability and efficiency.

AI Governance & Evaluation

Real-time monitoring, evaluation, and control systems for autonomous AI operations — tracking agent decisions, resource usage, and performance across multi-agent workflows.

Battle Rhythm

18 autonomous agents on coordinated cron schedules. The ecosystem manages itself.

            
            Always Running
          

*/2 min — GPU Thermal Guard

Throttle 75°C · Kill 80°C · Dual P40

*/5 min — Watchdog

47 services · Auto-restart · Circuit breaker

*/5 min — Reboot Monitor

Detects restarts · Re-launches services

Hourly — Auto-Commit

Preserves work every hour across repos

Hourly — Code Health

TODO drift · Test failures · Unpushed commits

Hourly — Activity Logger

Git activity + service state snapshot

Hourly — Article Agent

MuXD draft → Claude polish → Cloudflare deploy

*/2 hr — Feed Agent

Ecosystem status posts · 12 categories

            
            Daily Schedule
          

02:00 — Research Agent

State-of-the-art scan · 8 FS-RE layers · Brief

03:00 — Night Build

Snapshot → plan → build → test → deploy

06:15 — R&D Publisher

5-agent editorial → Wiki.js articles

06:30 — Daily Summary

Operations report · Git activity · Service health

06:45 — Podcast (Heart)

Script → Kokoro TTS → MP3 → Published

4x daily — Interview Prep

Flashcards from live ecosystem data → Scholar

12:00 — Noon Checkpoint

End protected hours · Resume operations

*/6 hr — Nextcloud Sync

Databases · Reports · Configs → Cloud backup

*/6 hr — Metrics Watchdog

Cross-validates counts across all domains

20:00 — Metrics Sync

Sync to WEBMASTER + Nextcloud + Broadcast

23:00 — Stats Updater

Update live numbers → Deploy to Cloudflare

Infrastructure Over Frameworks

When an agent fails at 3 AM, you want to tail a log file — not trace through a callback chain.

            ✗ LangChain / CrewAI
          

• Python classes, chains, graphs

• Framework runtime, event loops

• In-memory state management

• Framework-internal communication

• Step through chain logic to debug

            ✓ ARKONA
          

• Bash scripts + claude --print

• Cron schedules, OS process mgmt

• Filesystem + SQLite (survives crashes)

• Cron pipeline + message broker + MCP

• tail /tmp/agent.log — done

Each agent is a single file. Testing is bash agent.sh. Adding an agent is 5 lines of YAML. The same reason production databases use systemd, not a Python wrapper.

COMET — AI Governance Framework

Which tasks should humans own? Which should AI own? Who is accountable when things go wrong?

Upload Docs

SOPs, Org Charts

→

AI Analysis

Extract Roles/Tasks

→

Domain Wizard

5-Step Assessment

→

iPad Workshop

Facilitated Session

→

RACI Matrix

Human + AI Agent

Views

Full governance lifecycle

Standards

NIST, IEEE, ISO, MITRE

Delegation Levels

Human → Hybrid → AI

iPad

Workshop Mode

Admin + client devices

Walk into a room with iPads. Run a 2-hour facilitated session. Walk out with a standards-grounded RACI matrix showing exactly which tasks humans vs AI should own.

Let's Build Something

Currently exploring opportunities in agentic AI systems, AI safety, and applied AI research.

Get In Touch About Me

jhon.arango@IntrepidCyberSecEng.com
(667) 355-4069

Scan to save contact

Request Ecosystem Access

The ARKONA ecosystem is invite-only. Request an invite code to explore the platform.