jAIMSnet

Open-source AI Management Systems infrastructure — ISO/IEC 42001 aligned

j + AIMS (AI Management Systems) + net — Purpose-built for enterprise AI governance.

Project Overview

jAIMSnet is an AI management systems platform integrating multiple services for AI infrastructure management across ♾️ WeOwnNet 🌐 ecosystem.

Field	Value
Phase	Phase 1 🔴
Status	🔄 In Progress
Last Updated	2026-03-07
Cluster	jaimsnet-cluster · DigitalOcean DOKS · ATL1 · K8s 1.34.1
ISO Reference	ISO/IEC 42001:2023 AI Management Systems

---

Table of Contents

• Repository Structure
• Architecture
• Platform Status
• Infrastructure
• Domains
• Compliance Targets
• Directory Index
• Team
• Quick Start
• Cost Breakdown
• Links

---

Repository Structure

Monorepo (jaimsnet/jaimsnet) with structured directories for:

• docs/ (architecture, deployment guide, ADRs, runbooks)
• cluster/ (namespace definitions, ClusterIssuer)
• ingress/ (ingress-nginx, cert-manager)
• secrets/infisical/ (operator + sync CRDs)
• gateway/ (LiteLLM, Redis)
• observability/ (Langfuse, Prometheus, Grafana, Loki, Alloy, Mimir, Kuma)
• gpu/ (vLLM on MI300X)
• iac/ (OpenTofu, Ansible)
• gitops/ (ArgoCD, Watchtower)
• security/ (Kyverno, Trivy, Falco, CrowdSec, kube-bench, Docker Bench)
• compliance/ (NIST CSF, CIS Controls, ISO 27001, ISO 42001, CCM, FedRAMP)
• testing/ (integration, load, chaos)
• scripts/ (utility scripts)

Architecture

Internet → DO Load Balancer (129.212.240.75)
  → ingress-nginx
    → litellm.jAIMS.app  → LiteLLM     (gateway/)
    → langfuse.jAIMS.app → Langfuse    (observability/)
    → cert-manager (auto-TLS via Let's Encrypt)

LiteLLM → Redis            (gateway/redis)
LiteLLM → Langfuse         (callback / traces)
LiteLLM → OpenRouter / vLLM (providers)
LiteLLM → PostgreSQL       (spend tracking)
Langfuse → PostgreSQL      (traces + evals)

Infisical Cloud → Infisical Operator → K8s Secrets → Pods

Uptime Kuma (separate Droplet) → monitors all endpoints (kuma.jAIMS.app)

[Phase 2] vLLM on MI300X GPU Droplet (VPC) → LiteLLM backend
[Phase 2] Prometheus + Grafana + Loki + Alloy + Mimir
[Phase 2] Kyverno + Cilium Network Policies
[Phase 3] ArgoCD, Watchtower, Kyverno, Trivy, CrowdSec, Falco, kube-bench,
          docker-bench, Gitleaks, Syft, Grype, Ansible, full CI pipelines
[Phase 4] AI automation agents, customer-facing platform, k6 load testing,
          Litmus chaos engineering

---

Platform Status

Component	Tool	Phase	Status
Ingress controller	ingress-nginx	Phase 1 🔴	🔄 In Progress
TLS automation	cert-manager + Let's Encrypt	Phase 1 🔴	🔄 In Progress
AI gateway	LiteLLM	Phase 1 🔴	🔄 In Progress
Response caching	Redis	Phase 1 🔴	🔄 In Progress
LLM observability	Langfuse	Phase 1 🔴	🔄 In Progress
Secret management	Infisical Operator	Phase 1 🔴	🔄 In Progress
Endpoint monitoring	Uptime Kuma	Phase 1 🔴	🔄 In Progress
Infrastructure as Code	OpenTofu	Phase 1 🔴	🔄 In Progress
Metrics + dashboards	Prometheus + Grafana	Phase 2 🟠	📋 Planned
Log aggregation	Loki + Alloy	Phase 2 🟠	📋 Planned
Long-term metrics	Mimir	Phase 2 🟠	📋 Planned
GPU inference	vLLM on AMD MI300X	Phase 2 🟠	📋 Planned
Policy engine	Kyverno	Phase 2 🟠	📋 Planned
Network policies	Cilium	Phase 2 🟠	📋 Planned
GitOps (K8s)	ArgoCD	Phase 3 🟡	📋 Planned
GitOps (Docker)	Watchtower	Phase 3 🟡	📋 Planned
Image scanning	Trivy Operator	Phase 3 🟡	📋 Planned
Runtime security	Falco	Phase 3 🟡	📋 Planned
IDS + WAF	CrowdSec	Phase 3 🟡	📋 Planned
K8s benchmarking	kube-bench	Phase 3 🟡	📋 Planned
Docker benchmarking	docker-bench	Phase 3 🟡	📋 Planned
Secret scanning (CI)	Gitleaks	Phase 3 🟡	📋 Planned
SBOM generation	Syft + Grype	Phase 3 🟡	📋 Planned
Droplet automation	Ansible	Phase 3 🟡	📋 Planned
AI agents	TBD	Phase 4 🟢	📋 Planned
Load testing	k6 / Locust	Phase 4 🟢	📋 Planned
Chaos engineering	Litmus / Chaos Mesh	Phase 4 🟢	📋 Planned

---

Infrastructure

Resource	Spec	Cost (est.)	Status
DOKS Cluster	jaimsnet-cluster · ATL1 · K8s 1.34.1 · Premium AMD 2vCPU/8GiB · autoscale 1-2	~$24/mo	🔄 In Progress
DigitalOcean Load Balancer	129.212.240.75 · ATL1	~$12/mo	✅ Done
Managed PostgreSQL	ATL1 · port 25060 · sslmode=require	~$15/mo	✅ Done
Uptime Kuma Droplet	Basic Droplet · Docker	~$6/mo	📋 Planned
GPU Droplet (Phase 2)	AMD MI300X · ROCm · Docker	TBD	📋 Planned

---

Domains

Domain	Service	Status
litellm.jAIMS.app	LiteLLM AI Gateway	🔄 In Progress
langfuse.jAIMS.app	Langfuse LLM Observability	🔄 In Progress
kuma.jAIMS.app	Uptime Kuma Endpoint Monitor	📋 Planned
*.jaims.app	Wildcard (future services)	📋 Planned
jaims.network	Main platform domain	📋 Planned

---

Compliance Targets

Framework	Standard	Phase	Status
ISO/IEC 42001:2023	AI Management Systems	Phase 3 🟡	📋 Planned
ISO/IEC 27001:2022	Information Security Management	Phase 3 🟡	📋 Planned
SOC 2 Type II	Trust Service Criteria	Phase 3 🟡	📋 Planned
NIST CSF 2.0	Cybersecurity Framework	Phase 3 🟡	📋 Planned
CIS Controls v8	Security Best Practices	Phase 3 🟡	📋 Planned
CIS Benchmarks	K8s, Docker, Linux	Phase 3 🟡	📋 Planned
CSA CCM v4	Cloud Controls Matrix	Phase 3 🟡	📋 Planned
FedRAMP	US Federal Cloud Security	Phase 4 🟢	📋 Planned

See compliance/ for detailed framework mappings and evidence strategy.

---

Directory Index

Directory	Purpose	Phase
.github/	CI workflows, CODEOWNERS, PR template	Phase 3 🟡
docs/	Architecture, deployment guide, ADRs, runbooks	Phase 1 🔴
cluster/	DOKS cluster specs, namespace planning, network policies	Phase 1 🔴
ingress/	ingress-nginx + cert-manager (TLS automation)	Phase 1 🔴
secrets/	Infisical Operator + secret sync per namespace	Phase 1 🔴
gateway/	LiteLLM AI gateway + Redis caching	Phase 1 🔴
observability/	Langfuse (AI), Prometheus/Grafana/Loki (infra), Uptime Kuma	Phase 1–2
gpu/	vLLM on AMD MI300X GPU Droplet	Phase 2 🟠
iac/	OpenTofu (IaC) + Ansible (Droplet automation)	Phase 1–3
gitops/	ArgoCD (K8s GitOps) + Watchtower (Docker)	Phase 3 🟡
security/	Kyverno, Trivy, Falco, CrowdSec, kube-bench, docker-bench	Phase 2–3
compliance/	Frameworks, policies, evidence collection	Phase 3 🟡
testing/	Integration, load, and chaos testing	Phase 3–4
scripts/	Utility scripts (secret generation, backup, rotation)	Phase 1 🔴

---

Team

Member	Role	Responsibilities
@RMN	Platform + LLM Infrastructure	Gateway architecture, LiteLLM config, Langfuse, compliance strategy
@SHD	DevOps	Cluster operations, ingress, secrets, CI/CD, security tooling, Ansible
@LDC	Platform + Agentic AI	OpenTofu IaC, vLLM GPU, ArgoCD, Watchtower, AI automation
@GTM	Stakeholder	Requirements, compliance targets, customer platform, FedRAMP

---

Quick Start

Prerequisites: kubectl, helm (≥ 3.12), doctl (DigitalOcean CLI), tofu (OpenTofu)

# 1. Authenticate to the cluster
doctl kubernetes cluster kubeconfig save jaimsnet-cluster

# 2. Verify cluster context
kubectl config current-context
kubectl config use-context do-atl1-jaimsnet-cluster

# 3. Verify node pool
kubectl get nodes -o wide

# 4. Check all namespaces
kubectl get namespaces

# 5. Check running workloads
kubectl get pods -A

For full deployment steps see docs/deployment-guide.md. For operational runbooks see docs/runbooks/.

---

Cost Breakdown

Phase	Monthly Cost (est.)	Components
Phase 1 🔴	~$57/mo	DOKS (1 node) + LB + Managed PG + Kuma Droplet
Phase 2 🟠	~$112+/mo	Phase 1 + autoscale node + GPU Droplet
Phase 3 🟡	~$120+/mo	Phase 2 + additional tooling overhead
Phase 4 🟢	TBD	Phase 3 + customer platform infrastructure

Costs are estimates based on DigitalOcean ATL1 region pricing. Subject to change.

---

Links

Resource	URL
Platform domain	jaims.network
GitHub Organization	github.com/jAIMSnet
ISO/IEC 42001:2023	iso.org/standard/81230.html
LiteLLM docs	docs.litellm.ai
Langfuse docs	langfuse.com/docs
Infisical docs	infisical.com/docs
OpenTofu	opentofu.org